Hindawi Publishing Corporation
Journal of Aging Research
Volume 2012, Article ID 304014, 13 pages
doi:10.1155/2012/304014
Research Article
The Role of Lifestyle Behaviors on 20-Year Cognitive Decline
D. Cadar,1, 2 H. Pikhart,2 G. Mishra,3 A. Stephen,4 D. Kuh,1 and M. Richards1
1 MRC
Unit for Lifelong Health and Ageing, London WC1B 5JU, UK
of Population Health Sciences, University College London, London WC1E 6BT, UK
3 School of Population Health, University of Queensland, Herston, QLD 4006, Australia
4 MRC Human Nutrition Research, Elsie Widdowson Laboratory, 120 Fulbourn Road, Cambridge CB1 9NL, UK
2 Faculty
Correspondence should be addressed to D. Cadar, dorina.cadar.09@ucl.ac.uk and M. Richards, m.richards@nshd.mrc.ac.uk
Received 16 March 2012; Revised 27 May 2012; Accepted 3 July 2012
Academic Editor: Allison A. M. Bielak
Copyright © 2012 D. Cadar et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
This study examined the association between smoking, physical activity and dietary choice at 36 and 43 years, and change in these
lifestyle behaviors between these ages, and decline in verbal memory and visual search speed between 43 and 60–64 years in 1018
participants from MRC National Survey of Health and Development (NSHD, the British 1946 birth cohort). ANCOVA models
were adjusted for sex, social class of origin, childhood cognition, educational attainment, adult social class, and depression; then
the lifestyle behaviors were additionally mutually adjusted. Results showed that healthy dietary choice and physical activity were
associated, respectively, with slower memory and visual search speed decline over 20 years, with evidence that increasing physical
activity was important. Adopting positive health behaviors from early midlife may be beneficial in reducing the rate of cognitive
decline and ultimately reducing the risk of dementia.
1. Introduction
With an increase in the ageing population, the number of
older people affected by cognitive decline and dementia is
continually rising, causing a major public health impact
on individuals and governments around the world [1].
Despite major progress in understanding the neurobiology
of cognitive impairment and dementia, there are still no
clear determinants and complete causal models available for
explaining risks for this condition [2]. Substantial evidence
suggests that certain lifestyle behaviors (particularly smoking, sedentary lifestyle, and poor dietary choices) predict
faster cognitive decline [3–5], and higher risk of dementia
[6, 7], while physical, mental and social leisure activities are
found to be protective [8, 9]. A study of London civil servants
[10] highlights that the number and duration of unhealthy
behaviors are associated with subsequent cognitive function
in later life. Similar findings from the Suwon Longitudinal
Aging Study (SLAS) showed that a combination of multiple
positive lifestyle behaviors (such as nonsmoking, vegetable
consumption, and social activity) was associated with higher
cognitive ability [11]. However, since these behaviors tend to
cluster [12, 13], the extent to which apparent effects of one
behavior are attributable to (i.e., confounded by) another is
uncertain.
In addition, relatively little is known about the longitudinal effects of these behaviors on cognitive decline;
yet associations among multiple health behaviors place the
emphasis on longitudinal studies, since patterns of behaviors
tend to develop over decades, with implications for targeted
interventions to change the aggregate public health risk [14].
The life course approach to age-related diseases [15, 16]
provides an important opportunity to identify the nature and
timing of different environmental contributions to neuronal
damage and the risk of dementia across life [17]. The present
study therefore focuses on behavioral risk in early midlife, at
a stage of the life course when people are still more likely to
have control over modifiable risk and protective factors for
cognitive decline than in later life.
Although clinically significant cognitive decline and the
onset of dementia occur in older age (65+ years), it is
important to take in consideration early signs of cognitive
decline that appear in midlife. This is because conditions
such as dementia develop slowly and silently over the preceding decades [18–20]. Neuroimaging also show pathological
changes in midlife before the clinical signs of the disease
2
appear [21, 22]. Consistent with this, subtle cognitive decline
begins as least as early as the 5th decade [23, 24].
Risk and protective factors for health can exert their
most critical influences at different ages [25]. This is
acknowledged by the life course approach and the hypothesis
that positive lifestyle behaviors such as nonsmoking, being
physically active, and choosing healthier diets may protect
cognitive functioning and slow cognitive decline in later life.
Fratiglioni et al. identified key periods for potential risk and
protective factors [25]. Early life seems to be most critical
for the development of cognitive reserve (learning and
education) [26], when distal adverse influences (such as poor
childhood social circumstances) contribute to the risk of
adult disease or later life risk of dementia. Lifestyle behaviors,
including those that influence cardiovascular and metabolic
risk, become more influential in midlife, although some,
such as diet and physical activity, track back into childhood
[27, 28], whereas mental and physical activity patterns may
continue to moderate these risks into later life [29, 30].
The fact that the lifestyle behaviors are modifiable
implies that encouraging a healthy lifestyle may prevent or
ameliorate cognitive decline and underlying cerebrovascular
and cardiovascular risk factors [31]. Such interventions
should take into account the relative beneficial effect of
each independent behavior as well as their combined and
cumulative effect.
1.1. Present Study. The aim of the current study was to examine the role of individual and combined lifestyle behaviors
(smoking, physical activity, and dietary choice) on a 20-year
interval of cognitive decline. We used measures of memory
and psychomotor speed over this interval, which are sensitive
to decline associated with ageing and neurodegeneration.
The repeated measures of lifestyle behaviors at 36 and 43
years were used as predictors of cognitive change from 43 to
60–64 years. In addition to the independent and combined
effects of these behaviors, we also examined the cumulative
effects of these behaviors across early midlife and the changes
in these behaviors from one age to another.
2. Methods
2.1. Study Members. The Medical Research Council National
Survey of Health and Development (NSHD) originally
consisted of a socially stratified sample of 5362 children born
within marriage in one week in March 1946 in England,
Scotland, and Wales [32, 33]. This cohort has been followed
up prospectively 23 times, from birth onwards. In 1989, when
study members were aged 43 years, 3262 were successfully
contacted, of whom 3004 have completed both cognitive
tests. Of these, 1911 had successfully completed the most
recent cognitive assessment, conducted between 2006 and
2011 when survey members were aged 60 to 64 years (henceforth 60+ years). This data collection began with a postal
questionnaire sent to the target 3163 sample [33]. This was
followed by an invitation to an assessment by trained nurses
at one of six clinical research facilities (based in Cardiff,
Birmingham, Edinburgh, London (at UCL and St Thomas’
hospitals), and Manchester), or, if preferred, at a home
Journal of Aging Research
Original sample 1946
N = 5362
Contacted in 1989
Age 43
N = 3004 (63%)
Followed up in 2009–2011
cognitive assessments
Age 60+
N = 1911
Not contacted in 1989
N = 2358 (37%)
(previously refused 12%,
living abroad 11%, untraced 5%,
died 9%)
Not missing data on the 2 follow-up measures for each of the 3 lifestyle
behaviors (smoking, physical activity, and diet)
N = 1201
Not missing data on all the covariates
Final study sample with complete data
N = 1018
Figure 1: Flow chart of lifestyle behaviors, cognitive functioning
and covariates data follow up (considered since baseline cognitionat age 43) in the Medical Research Council National Survey of
Health and Development.
visit. This was supplemented by a further questionnaire
sent to those who agreed to a clinic or home visit. These
questionnaires, along with additional questions asked by the
nurses, updated information on general health, household
composition, family structure, socioeconomic status, daily
function, life events, and lifestyle. Contact was not attempted
for 2198 (718 deaths, 567 living abroad, 594 prior refusals,
and 320 permanently lost). From the current respondents,
a total of 1,018 study members had nonmissing data for all
variables incorporated in these analyses (see Figure 1).
2.2. Ethical Approval. The study protocol received ethical
approval from the Greater Manchester Local Research Ethics
Committee for the five English sites; from Scotland Research
Ethics Committee for the data collection taking place in
Edinburgh. Written informed consent was obtained from the
study member at each stage of data collection.
2.3. Cognitive Function. Cognitive functioning was measured
at 43 and 60+ years using two tests: verbal memory and speed
and concentration. The verbal memory test consisted of a 15item word learning task devised by the NSHD research team.
Each word was shown for two seconds. When all 15 words
were shown, the study member was asked to write down as
many of these as possible, in any order. The total number
Journal of Aging Research
of words correctly recalled over three identical trials was
summed to provide an overall score for short-term verbal
memory (maximum score = 45). This was followed by a letter
search test (see below), after which an uncued delayed free
recall trial was administered.
Speed and concentration were assessed with a letter
search task in which participants were required to cross out
the letters P and W, randomly embedded within a page of
other letters, as quickly as possible within 1 minute. At age
43 years the total number of letters to be searched was 450,
and at age 60+ years it was 600. Scores were computed as total
number of letters searched.
2.4. Lifestyle Behaviors. Data on lifestyle behaviors were
extracted from questionnaires, interview-based prospective
information, and diet diaries completed in early midlife (36
and 43 years).
2.5. Smoking. Interview-based prospective information on
cigarette smoking frequency was obtained at age 36 and 43
years. Smoking frequency was categorized at each age as
0 (nonsmoker), 1 to 20 (light smoker), or more than 20
cigarettes per day (heavy smokers).
A midlife smoking score was derived by assigning those
classified as nonsmokers a value of 0, those as light smokers
a value of 1, and those as heavy smokers a value of 2 at
each age (36 and 43 years) and then summing the values
for the 2 ages whereby an individual with a midlife smoking
score of 0 was categorized as nonsmoker, an individual with a
midlife smoking score of 1 to 3 was categorized as moderate
smoker, and the remaining individuals were categorized as
heavy smokers (sum equal 4).
A score for change in smoking behavior was also derived
by assigning those classified as nonsmokers at each age a
value of 0; those reporting an increase in smoking a score
of 1; those who reported a decrease in smoking a score of 2;
those who were constantly moderate or heavy smokers across
early midlife a score of 3.
2.6. Physical Activity. Physical activity levels were ascertained
at ages 36 and 43 years during interviews at the study participants’ home. Questions about physical activity at age 36 years
were based on the Minnesota leisure time physical activity
questionnaire [27, 34]. The questions addressed engagement
in sports and recreational activities in the previous month,
utilizing a checklist of 27 different leisure time activities. At
age 43 years, participation in any sports, vigorous leisure
activities, or exercise was reported, although this was based
on answers to an open-ended question, rather than the above
checklist. However, the monthly frequency of these activities
was also reported, enabling a similar categorization to the
36 year measure. At each age, participants were categorized
as inactive (reported no participation); moderately active
(participated in relevant activities 1–4 times in the previous
month at age 36 years, and per month at age 43 years); or
most active (participated in relevant activities five or more
times in the previous month at age 36 years and per month
at age 43 years) [35].
3
A total physical activity score was derived by assigning
those classified as inactive a value of 0, those as moderately
active a value of 1, and those as most active a value of 2 at
each age (36 and 43 years) and then summing the values
for the two ages. This was categorized as inactive (score 0),
moderately active (score 1 to 3), and most active (score 4).
A score for change in physical activity behavior, was
derived in a similar fashion as for smoking, by assigning
those classified as inactive at each age a value of 0; those
reporting an increase in activity a score of 1; those who
reported a decrease in their physical activity levels a score of
2; those whose physical activity levels were moderate or most
active at both ages a score of 3.
2.7. Dietary Choice. Dietary intake was assessed by a fiveday diary [36] at both 36 and 43 years. All food and drinks
consumed both at and away from home were recorded in
the diaries, including brand names of food products, food
preparation methods, and recipes used. Participants were
asked to record the amount eaten in household measures,
with guidance notes and photography provided in the diary
to assist in estimating portion size [37]. From this information, an overall score representing level of healthy food choice
was derived, by summing scores for four separate criteria:
(1) consumption of breakfast (a score of 0 representing no
consumption to 1 some days and 2 all days); (2) type of milk
(from 0 whole only to 3 skim milk only); (3) type of bread
(from 0 white only to 4 wholemeal only); (4) number of
daily portions of fruit and vegetables (from 0 to maximum
5 portions per day); a dietary reference score representing
the percentage of energy from fat, carbohydrates, and protein
(scores from 1-highest to 5-lowest percentages less than
30% energy). The total score was subject to a median split
(median = 10 at age 36 years and median = 11 at age 43 years;
minimum score 0 and maximum 19) to represent low versus
high (energy dense/nutrient poor versus healthier) dietary
choice. The range of dietary scores varied from a minimum
of 5 to a maximum of 19; M = 10.22, SD = 1.84 at age 36;
M = 11.42, SD = 2.27 at age 43).
A midlife dietary choice score was also derived by
assigning those classified as making poorer choices for their
diet a value of 0 and those making healthier choices a value
of 1 at each age (36 and 43 years) and then summing up the
values for the 2 ages. Midlife dietary choice was categorized
as making poorer (low) choices at both ages (score 0) and
making healthier dietary choices (high) at least at one age
(score 1 or 2).
A score for change in diet was also derived. Those
classified as choosing a low-quality diet at each age were
assigned a value of 0; those reporting an increase from a lowquality to a high-quality diet a score of 1; those who changed
to a lower-quality diet a score of 2; those who constantly
maintain a high-quality diet were assigned a score of 3.
2.8. Covariates. Based on previous findings, the following variables were treated as potential confounders: sex;
father social class; childhood cognitive ability; educational
attainment; midlife household occupational social class and
depression [14, 38–40].
4
2.9. Occupational Social Class of Origin. Occupational social
class of origin was represented by father’s social class when
participants were aged 11 years, or if this was unknown,
at age 4 or 15 years. This was classified as professional
managerial intermediate, skilled nonmanual, skilled manual,
semi-skilled manual, or unskilled, according to the UK
Registrar General’s Classification of Occupations [41].
2.10. Prior Cognitive Ability. Childhood cognitive ability at
age eight years was represented as the sum of four tests
of verbal and nonverbal ability devised by the National
Foundation for Educational Research [42]. These tests were
(1) reading comprehension (selecting appropriate words to
complete 35 sentences); (2) word reading (ability to read and
pronounce 50 words); (3) vocabulary (ability to explain the
meaning of 50 words); (4) picture intelligence, consisting of
a 60-item nonverbal reasoning test.
2.11. Educational Attainment. The highest educational qualification achieved by age 26 years was dichotomized into
those with advanced (“A level”, taken during the final year of
secondary/high school) or higher (university or equivalent)
qualifications, versus those below this level.
2.12. Occupational Social Class. Household midlife occupational social class was used at each time of the health
behavior (36 or 43 years), or the latest age for longitudinal
behavioral measures (see below). This was coded according
to the Registrar General (as for social class of origin).
2.13. Depression and Anxiety Symptoms. Frequency and
severity of common symptoms of depression and anxiety
were assessed by the Psychiatric Symptom Frequency scale
(PSF) [43] at age 43 years, and with the 28-item General
Health Questionnaire (GHQ) [44] at age 60+.
2.14. Statistical Analyses. Multivariable ANCOVA models
were used to test associations between smoking, physical
activity, and diet at each age, entered as categorical variables
and cognitive decline, used as continuous variables. First,
we tested the association between each lifestyle behavior at
36 and 43 years and change in both cognitive outcomes
(verbal memory and letter search speed); then we used the
cumulative behavior scores for both these ages; finally we
tested associations between change in behaviors between
these ages and change in the cognitive outcomes between
ages 43 and 60+ years. In order to reduce the effect of
regression to the mean and because of the difference in size
of the letter search matrix at ages 43 and 60+ years, we
used conditional models of change by adjusting cognitive
scores at age 60+ years for their corresponding score at
43 years [23]. Positive coefficients represent a slower rate
of decline, and negative coefficients represented a faster
rate of decline. Model 1 adjusted raw associations for sex
(since there was no evidence of sex x lifestyle behavior
interactions on the outcomes), social class of origin, prior
cognitive ability at age 8 years, educational attainment,
occupational class, and symptoms of anxiety and depression.
Model 2 included these covariates but additionally tested
Journal of Aging Research
the specificity of lifestyle behaviors by mutually adjusting
each for the other behaviors. For cumulative effects, model
2 was further adjusted for the cumulative scores of the
additional behaviour; for change in behaviour from 36 to
43, model 2 was mutually adjusted for the change in other
lifestyle behaviours. All analyses were based on the sample
with complete data on cognitive tests at the 2 time points,
lifestyle behaviors at the 2 time points, and all the selected
covariates.
3. Results
3.1. Sample Description and Missing Data. Participants with
missing scores on memory and visual search at age 43 were
more likely to be men, to have less than advanced educational
attainment, to be in a manual occupation in midlife, and
to have lower cognitive capability scores at age 8 years (all
P < 0.001). A similar pattern was observed for those without
cognitive scores at age 60+. Those with missing scores on
memory at age 43 had a marginally higher total anxiety and
depression score at the same age than those who underwent
memory testing (P = 0.058). Those who did not complete
data on lifestyle behaviors at both 36 and 43 years had
lower cognitive scores for both verbal memory (P < 0.001)
and visual search (P = 0.076; P = 0.010) at age 43 and
60+, were more likely to belong to a manual occupation
of social class of origin (P < 0.001), and had less than
advanced levels of educational attainment by age 26 years
(P < 0.001).
Participants included in the current analysis were mostly
nonsmokers at either 36 (76.6%) or 43 years (80.1%) and at
least moderately active at ages 36 (68.1%), although less so
at 43 years (53.2%), but had a similar quality of diet at both
36 and 43 years (55.4% and 51.1%, respectively, were in the
lower median split).
Table 1 shows means for the total verbal memory and
visual search speed scores at age 60+, by the three lifestyle
behaviors (at 36 and 43 years, and the cumulative and change
scores for these ages), by sex, father’s social class, educational
attainment, and midlife social class and depression of the
1018 participants. All three lifestyle behaviors (smoking,
physical activity, and diet) at age 36, 43 years, as cumulative
scores or change in behavior were strongly associated with
verbal memory at age 60+. The effect was monotonic,
with those who did not smoke or decreased their level of
smoking; those who were most physically active or increased
their level of activity or had a healthier or improved diet,
having better memory scores at age 60+. For visual search
speed at age 60+ the general trend was for slower function
with heavy and increased smoking across early midlife; a
positive association for high and increased physical activity;
a positive association for consistently high and increased diet
quality.
3.2. Association between Lifestyle Behaviors and Memory
Decline. Table 2 shows results for the ANCOVAs for each
health behavior (at 36 and 43 years, the cumulative and
change scores for these ages) and memory decline from
43 to 60+ years. Trends can be observed for associations
Journal of Aging Research
5
Table 1: Characteristics and mean scores for memory and search speed at age 60+ years, by lifestyle behaviour profiles and dichotomous
covariates.
Variable
Smoking at age:
36 years
Non-smoker
1–20 c/day
20+ cig/day
43 years
Non-smoker
1–20 c/day
20+ cig/day
Midlife score
Non-smoker
Moderate smoker
Heavy smoker
Change in smoking
Non-smoker
Increase
Decrease
Constant med/heavy
Physical activity at age:
36 years
Inactive
Moderate active
Most active
43 years
Inactive
Moderate active
Most active
Midlife score
Inactive
Moderate active
Most active
Change in physical activity:
Inactive
Increase
Decrease
Constant active
Diet at age:
36 years
Low quality
High quality
43 years
Low quality
High quality
Midlife score
Low quality
High quality
Change in diet:
Constant low quality
Increase
Decrease
Constant high
N
Memory at age 60+ years
P
Search speed at age 60+ years
P
780
200
38
25.00 (5.88)
24.18 (6.37)
21.23 (7.33)
<0.001
273.47 (73.13)
266.7 (69.88)
233.26 (59.33)
0.002
816
159
43
25.05 (5.95)
23.99 (6.05)
20.72 (7.05)
<0.001
274.45 (72.55)
258.60 (68.67)
242.83 (73.49)
0.001
744
252
22
25.05 (5.90)
24.07 (6.30)
20.18 (7.25)
<0.001
274.51 (73.69)
263.48 (67.50)
221.81 (59.74)
<0.001
744
55
79
140
25.05 (5.90)
22.89 (5.91)
24.77 (6.51)
23.52 (6.60)
0.005
274.51 (73.69)
254.63 (65.05)
272.29 (60.10)
255.44 (72.33)
0.011
324
275
419
23.82 (5.83)
24.61 (5.98)
25.44 (6.24)
<0.001
264.87 (72.81)
270.04 (68.14)
275.49 (74.58)
0.138
475
251
292
23.85 (6.13)
25.15 (6.05)
25.69 (5.84)
<0.001
263.83 (69.38)
267.10 (69.13)
284.76 (78.01)
<0.001
221
608
189
23.37 (6.01)
24.81 (5.98)
25.90 (6.20)
<0.001
259.08 (74.61)
270.87 (68.78)
283.43 (79.00)
0.003
221
164
354
279
23.37 (6.01)
25.03 (5.35)
24.61 (6.30)
25.68 (6.08)
<0.001
259.08 (74.61)
282.39 (74.40)
266.84 (66.91)
277.72 (74.71)
0.003
564
454
24.01 (6.19)
25.55 (5.84)
<0.001
267.26 (69.12)
274.85 (76.14)
0.096
521
497
23.72 (6.13)
25.72 (5.86)
<0.001
267.24 (69.30)
274.21 (75.41)
0.124
234
784
22.51 (5.90)
25.34 (5.99)
<0.001
260.02 (67.96)
273.81 (73.41)
0.010
234
330
94
360
22.55 (5.90)
25.06 (6.18)
23.98 (5.44)
25.96 (5.88)
<0.001
260.02 (67.96)
272.39 (69.57)
263.80 (71.44)
277.73 (77.16)
0.022
6
Journal of Aging Research
Table 1: Continued.
Variable
Sex:
Male
Female
Father social class
Non-manual
Manual
Education:
<advanced
≥advanced
Midlife social class at age:
36 years
Non-manual
Manual
43 years
Non-manual
Manual
Depression at age:
43 years
No symptoms
Depressive symptoms
60+ years
No symptoms
Depressive symptoms
N
Memory at age 60+ years
P
Search speed at age 60+ years
P
490
528
23.72 (5.94)
25.61 (6.07)
<0.001
265.36 (71.84)
275.54 (72.63)
0.024
507
511
26.28 (5.90)
23.12 (5.84)
<0.001
278.69 (75.64)
262.66 (68.16)
<0.001
567
451
22.77 (5.66)
27.12 (5.71)
<0.001
263.03 (70.09)
280.21 (74.17)
<0.001
617
401
26.24 (5.79)
22.33 (5.75)
<0.001
280.36 (71.29)
255.69 (71.62)
<0.001
696
322
26.08 (5.79)
21.72 (5.60)
<0.001
276.31 (71.73)
258.38 (72.42)
<0.001
966
52
24.80 (6.05)
22.98 (6.47)
0.036
271.08 (72.92)
262.46 (61.78)
0.403
805
171
24.85 (6.09)
24.19 (5.74)
0.194
270.03 (73.61)
273.87 (69.23)
0.531
between heavy smoking at age 43 and for increase in smoking
consumption and faster memory decline at all stages of
model adjustment, but these trends were not significant at
the α = 0.05 level. Directions of associations between physical
activity and memory decline were inconsistent across the
various categories of this behavior, but none were significant.
On the other hand a consistently healthy dietary choice
at 36 and 43 years was associated with slower memory
decline, although none of the dietary change categories were
associated with this outcome.
3.3. Association between Lifestyle Behaviors and Visual Search
Speed Decline. Table 3 shows results for the ANCOVAs for
each health behavior (at 36 and 43 years, and the cumulative
and change scores for these ages) and decline in visual search
speed from 43 to 60+ years. Heavy smoking at 43 years was
associated with faster decline in search speed compared to
nonsmokers after adjustment for the covariates in Model
1, but not at 36 years or at both these ages (the latter
an effect of reduced statistic power) and not at 43 years
after mutual adjustment for the other health behaviors. Nor
were there associations between change in smoking and this
outcome. On the other hand high physical activity at 43 years
was associated with slower search speed decline, after full
adjustment. This was also the case for those most active at
both 36 and 43 years and for those who increased their level
of activity between these ages. There were no associations at
the α = 0.05 level between any measure of dietary choice and
rate of search speed decline.
4. Discussion
The principal aim of this study was to test associations
between lifestyle behaviors in early midlife and cognitive
decline over 20 years. Key findings were that a consistently
healthy dietary choice was associated with slower memory
decline, and that consistently high or increasing physical
activity from early midlife to midlife was associated with
slower visual search speed decline, independently of each
other lifestyle behavior and of social class of origin, childhood cognition, educational attainment, adult social class,
and symptoms of anxiety and depression. Smoking was not
associated with either cognitive outcome. It should be noted
that the current findings for dietary choice and physical
activity were not always consistent at different ages across
midlife, compared to effects of the cumulative scores and
change in behavior between the 2 time points. It will be
important to seek replication elsewhere before these findings
can be treated as authoritative.
4.1. Strengths and Limitations. Strengths of this study include
use of a nationally representative sample; availability of a
wide range of prospectively-obtained potential confounders,
including childhood cognition to rule out selection by prior
ability; a detailed assessment of lifestyle behaviors in midlife,
including 5-day diet diaries; estimation of cumulative and
change effects with two measures of health behaviors;
availability of cognitive outcomes that are age and morbidity
sensitive [45, 46]; a 20-year interval for capturing decline in
Journal of Aging Research
7
Table 2: Regression coefficients (95% confidence intervals) representing rate of decline in verbal memory from 43 to 60+ years per level
increase in each health behavior at 36 and 43 years and per cumulative and change in health behavior scores.
N (%)
Smoking at age:
36 y
Non-smoking
1–20 c/day
20+ cig/day
P value
43 y
Non-smoking
1–20 c/day
20+ cig/day
P value
Midlife score:
Non-smoker
Moderate smoker
Heavy smokers
P value
Change in smoking:
Non-smoker
Increase
Decrease
Constant med/heavy
P value
Physical activity at age:
36 y
Inactive
Moderately active
Most active
P value
43 y
Inactive
Moderately active
Most active
P value
Midlife score:
Inactive
Moderate
Most active
P value
Change in activity:
Inactive
Increase
Decrease
Constant mod/active
P value
Diet at age:
36 y
Low-quality diet
High-quality diet
P value
43 y
Low-quality diet
High-quality diet
P value
780 (76.6)
200 (19.6)
38 (3.7)
816 (80.1)
159 (15.6)
43 (4.2)
744 (73.0)
252 (24.7)
22 (2.1)
744 (73.0)
55 (5.4)
79 (7.7)
140 (13.7)
324 (31.8)
275 (27.0)
419 (41.1)
475 (46.6)
251 (24.6)
292 (28.6)
221 (21.7)
608 (59.7)
189 (18.5)
221 (21.7)
164 (16.1)
354 (34.7)
279 (27.4)
Verbal memory decline 43 to 60+ (95% CI)
Model 1
Model 2
0.00
0.00
−0.16 (−0.85, 0.53)
−0.22 (−1.67, 1.22)
−0.12 (−0.81, 0.57)
−0.17 (−1.62, 1.28)
0.610
0.701
0.00
0.00
−0.08 (−0.83, 0.67)
−1.38 (−2.75, −0.01)
−0.07 (−0.83, 0.69)
−1.24 (−2.61, 0.11)
0.124
0.165
0.00
0.00
−0.20 (−0.83, 0.43)
−0.22 (−2.10, 1.66)
−0.12 (−0.76, 0.51)
−0.01 (−1.90, 1.87)
0.535
0.750
0.00
0.00
−1.11 (−2.32, 0.09)
0.10 (−0.91, 1.12)
−0.03 (−0.83, 0.77)
−1.00 (−2.21, 0.21)
0.09 (−0.92, 1.11)
0.04 (−0.78, 0.86)
0.363
0.473
0.00
0.00
−0.06 (−0.76, 0.64)
0.21 (−0.42, 0.86)
−0.07 (−0.78, 0.63)
0.17 (−0.47, 0.82)
0.484
0.571
0.00
0.06 (−0.61, 0.75)
−0.06 (−0.72, 0.59)
0.878
0.00
0.02 (−0.65, 0.71)
−0.12 (−0.79, 0.54)
0.731
0.00
0.00
−0.13 (−0.82, 0.54)
−0.16 (−1.05, 0.71)
−0.23 (−0.91, 0.45)
−0.31 (−1.20, 0.56)
0.697
0.472
0.00
0.00
−0.35 (−1.25, 0.54)
−0.10 (−0.84, 0.63)
−0.06 (−0.87, 0.73)
−0.47 (−1.37, 0.43)
−0.18 (−0.92, 0.55)
−0.17 (−0.98, 0.63)
0.871
0.768
564 (55.4)
454 (44.6)
0.00
0.38 (−0.17, 0.93)
0.176
0.00
0.35 (−0.20, 0.91)
0.212
521 (51.1)
497 (48.8)
0.00
0.20 (−0.34, 0.75)
0.468
0.00
0.17 (−0.38, 0.73)
0.549
8
Journal of Aging Research
Table 2: Continued.
N (%)
Midlife score:
Low-quality diet
High-quality diet
P value
Change in diet:
Low-quality
Increase in quality
Decrease in quality
Constant high-quality
P value
Verbal memory decline 43 to 60+ (95% CI)
Model 1
Model 2
234 (22.9)
784 (77.0)
0.00
0.71 (0.04, 1.38)
0.037
0.00
0.70 (0.02, 1.38)
0.043
234 (22.9)
330 (32.4)
94 (9.2)
360 (35.3)
0.00
0.65 (−0.08, 1.40)
0.95 (−0.09, 2.00)
0.78 (0.02, 1.53)
0.146
0.00
0.65 (−0.10, 1.41)
0.91 (−0.14, 1.97)
0.76 (0.00, 1.54)
0.180
Model 1: model adjusted for sex, childhood social class, childhood cognition at age 8, SEP (own occupation and education), and depression.
Model 2: model 1 plus other lifestyle behaviors.
these measures upto an age when decline may begin to have
functional consequences. In addition, while there are many
reports examining associations with single lifestyle behaviors
[47–49], very little work has focused on the combined
influence of these behaviors on cognitive functioning [10, 11]
and none at all to our knowledge on cognitive decline.
Furthermore, we believe that the long interval between repeat
administrations of the cognitive tests would have minimized
potential practice effects.
We should also highlight a number of limitations. First,
there was a disproportional loss of those who were less
advantaged, in terms of lower childhood cognitive ability,
education, and SES, although we have no reason to suspect
that this affected the pattern of results observed. Second, all
information on lifestyle behaviors was dependent on selfreport and was not validated by independent measures, for
example, cotinine in the case of smoking. Third, there is the
possibility of regression to the mean effect when analyzing
cognitive ability data over a long period of time, for example,
one may assume that some of those who had high cognitive
ability score at age 43 may have lower scores at age 63 years
and vice versa. We have tried to reduce this effect by using the
conditional model of change [50].
In regard to previous findings in NSHD, the inverse
association between physical activity and memory decline
from 43 to 53 years reported by Richards et al. [51] was
not seen here for 20-year decline. This may be because the
previous study had additional specificity through adjusting
for non-physical spare-time activities, which would have
been too cumbersome in the context of multiple lifestyle
behaviors in the present study, or because of the longer
period of cognitive change. The associations between physical activity and slower decline in visual search speed and
between healthy dietary choice and memory are new findings
and were not previously tested in this cohort; in the former
case physical activity was not investigated in relation to
search speed in the previous study; in the latter case midlife
cognition has not previously been studied in relation to
diet in this cohort. On the other hand the associations
between heavy smoking at age 43 and faster memory decline
previously reported between ages 43 and 53 years [40] were
not replicated here with the 20-year period of cognitive
change from 43 to 60+. The loss of the cumulative midlife
heavy smoking-memory decline association may be due low
statistical power resulting from the relatively high odds of
morbidity and premature mortality in this subgroup (135
study members smoking more than 20 cigarettes per day at
age 43 were represented in the previous study, compared to
22 in the present study).
In relation to other cohorts, Sabia et al. found an effect
of sex on the association between smoking and cognitive
decline in a study of London civil servants. Their results
showed that men who smoked showed faster decline than
nonsmoking men over a 10-year period, after adjusting for
the effects of heart disease, stroke, and lung function on
mental abilities, while for women there were no differences
in cognitive scores over the same time period. This could
be related to a lower number of female participants in
contrast to males in the Whitehall II study [52]. In relation
to physical activity, leisure-time physical activity at least
twice a week in midlife was associated with reduced risk of
memory decline in the Cardiovascular Risk Factors, Aging
and Incidence of Dementia (CAIDE) study after adjustment
for age, sex, education, follow-up time, locomotor disorders,
APOE genotype, vascular disorders, smoking, and alcohol
consumption [53]. Similarly, in The Mayo Clinic Study of
Aging, moderate exercise in midlife or late life was associated
with reduced odds of Mild Cognitive Impairment (MCI)
[54]. In contrast, results from the Chicago Health and Aging
Project reported that physical activity conducted within
2 weeks of the date of baseline cognitive assessment was
not associated with risk of cognitive decline in an older
population [55].
Our finding that maintained healthy dietary choice was
associated with a slower rate of memory decline is consistent with the results of a recent systematic review, which
highlighted that a diet high in saturated fat represents an
increased risk of cognitive decline and subsequent dementia
[6]. The emphasis on identifying specific nutrients associated
with cognitive ability in later life, such as antioxidants
(vitamin C, E, carotenoids, and polyphenols), minerals,
and dietary lipids (total, trans, and saturated mono-and
Journal of Aging Research
9
Table 3: Regression coefficients (95% confidence intervals) representing rate of decline in visual search from 43 to 60+ years per level
increase in each health behavior at 36 and 43 years and per cumulative and change in health behavior scores.
N (%)
Smoking at age:
36 y
Non-smoking
1–20 c/day
20+ cig/day
P value
43 y
Non-smoking
1–20 c/day
20+ cig/day
P value
Midlife score:
Non-smoker
Moderate smoker
Heavy smokers
P value
Change in smoking:
Non-smoker
Increase
Decrease
Constant med/heavy
P value
Physical activity at age:
36 y
Inactive
Moderately active
Most active
P value
43 y
Inactive
Moderately active
Most active
P value
Midlife score:
Inactive
Moderate
Most active
P value
Change in activity:
Inactive
Increase
Decrease
Constant mod/active
P value
Diet at age:
36 y
Low-quality diet
High-quality diet
P value
43 y
Low-quality diet
High-quality diet
P value
780 (76.6)
200 (19.6)
38 (3.7)
816 (80.1)
159 (15.6)
43 (4.2)
744 (73.0)
252 (24.7)
22 (2.1)
744 (73.0)
55 (5.4)
79 (7.7)
140 (13.7)
324 (31.8)
275 (27.0)
419 (41.1)
475 (46.6)
251 (24.6)
292 (28.6)
Visual search decline 43 to 60+ (95% CI)
Model 1
Model 2
0.00
0.00
−1.90 (−11.77, 7.96)
−18.70 (−39.27, 1.86)
−1.79 (−11.68, 8.09)
−19.44 (−40.05, 1.17)
0.156
0.148
0.00
0.00
−4.19 (−14.92, 6.53)
−19.54 (−38.97, −0.12)
−2.73 (−13.55, 8.08)
−18.52 (−38.00, 0.94)
0.060
0.103
0.00
0.00
−4.85 (−13.93, 4.21)
−24.80 (−51.61, 2.00)
−4.82 (−13.93, 4.28)
−25.46 (−52.34, 1.42)
0.078
0.077
0.00
0.00
−13.08 (−30.17, 4.00)
−2.70 (−17.06, 11.64)
−6.71 (−18.09, 4.67)
−12.78 (−30.02, 4.45)
−2.65 (−17.09, 11.78)
−5.82 (−17.44, 5.79)
0.374
0.455
0.00
2.69 (−7.29, 12.68)
8.00 (−1.12, 17.13)
0.081
0.00
3.20 (−6.80, 13.20)
8.47 (−0.70, 17.65)
0.067
0.00
0.00
−3.93 (−13.57, 5.69)
−4.37 (−14.03, 5.28)
11.84 (2.55, 21.12)
0.022
11.18 (1.78, 20.58)
0.034
221 (21.7)
608 (59.7)
189 (18.5)
0.00
8.32 (−1.33, 17.97)
16.71 (4.32, 29.10)
0.008
0.00
8.76 (−0.95, 18.47)
17.00 (4.50, 29.51)
0.008
221 (21.7)
164 (16.1)
354 (34.7)
279 (27.4)
0.00
14.45 (1.83, 27.07)
7.58 (−2.87, 18.03)
10.51 (−0.74, 21.77)
0.127
0.00
13.78 (1.05, 26.50)
7.80 (−2.72, 18.34)
10.46 (−0.92, 21.84)
0.157
564 (55.4)
454 (44.6)
521 (51.1)
497 (48.8)
0.00
0.00
−0.52 (−8.39, 7.34)
−1.74 (−9.67, 6.19)
0.896
0.667
0.00
0.59 (−7.23, 8.42)
0.881
−1.03 (−8.91, 6.84)
0.00
0.797
10
Journal of Aging Research
Table 3: Continued.
N (%)
Midlife score:
Low-quality diet
High-quality diet
P value
Change in diet:
Low-quality
Increase in quality
Decrease in quality
Constant high-quality
P value
234 (22.9)
784 (77.0)
234 (22.9)
330 (32.4)
94 (9.2)
360 (35.3)
Visual search decline 43 to 60+ (95% CI)
Model 1
Model 2
0.00
1.59 (−7.89, 11.09)
0.741
−0.64 (−10.28, 9.00)
0.00
0.00
1.53 (−9.04, 12.12)
−1.23 (−16.11, 13.64)
1.63 (−9.09, 12.36)
0.972
−0.69 (−11.45, 10.06)
−3.22 (−18.18, 11.74)
−1.39 (−12.31, 9.57)
0.896
0.00
0.978
Model 1: model adjusted for sex, childhood social class, childhood cognition at age 8, SEP (own occupation and education), and depression.
Model 2: model 1 plus other lifestyle behaviors.
polyunsaturated fats) [56], is now giving way to studies of
global diet quality indices, for guidance in modeling dietary
risk in relation to cognitive performance [57, 58]. In this context general recommendations are made for high fruit and
vegetable consumption and moderation of high-glycemic
index foods as cardioprotective, with secondary benefits to
cognitive ageing [59]. The reason for the specificity of the
diet memory association is unclear. Healthy dietary choice
is protective of dementia [60], and evidence shows that
dementia is predicted by memory impairment in particular
[61, 62]. However, physical exercise, which was specifically
associated with search speed, is also protective of dementia
[63]. As already noted, however, the apparent loss of the
exercise memory association reported by Richards et al.
(2003) may be a consequence of not adjusting for non
physical leisure activities here [51].
Our results for diet differ to those of The Nutrition
et Cognition (NutCog) study, where diet quality was not
independently associated with cognitive change over 3year period. However, cognitive decline in this study was
measured over a much shorter interval, although this study
did report an association between diet quality and risk
factors for nutrition-related chronic diseases, which are also
considered to be risk factors for cognitive decline [56].
There are several plausible biological mechanisms underlying associations between physical activity and cognitive
decline. Physical activity reduces cardiovascular risk [64],
increases cerebral perfusion, and facilitates neurogenesis
[65–67]. In contrast, impaired blood flow in the midbrain
[68, 69] is a risk factor for subsequent cognitive impairment
and dementia. Of interest, physical activity was specifically
associated with a slower decline in psychomotor speed,
which is consistent with evidence that highly fit individuals
respond faster to stimuli [70].
4.2. Conclusions and Implications. In conclusion, our results
support evidence that physical activity and healthy dietary
choice are protective of aspects of cognitive ageing. These
mutually adjusted associations were observed over a 20year period, in analyses additionally controlling for socioeconomic status, sex, educational attainment, prior cognitive
ability, and symptoms of anxiety and depression. Further
work on interactions between lifestyle behaviors is recommended. For example, evidence suggests that smokers
have poorer dietary choices than nonsmokers [71]; the
Mediterranean diet only appears to be an effective protective
factor for Alzheimer’s disease in those who also exercise [72].
Overall, however, and in view of the enormous financial
and societal burden of neurodegenerative diseases, public
health interventions based on modifiable lifestyle behaviors
across the life course represent high level priorities around
the world.
Acknowledgments
D. Cadar, A. Stephen, D. Kuh, and M. Richards are supported by the UK Medical Research Council. H. Pikhart
is supported by the Higher Education Funding Council
for England (HEFCE). G.Mishra receives support from the
Australian National Health and Medical Research Council
(APP1000986). No financial disclosures were reported by the
authors of this paper. Financial support for this study was
provided by the UK Medical Research Council. The authors
are grateful to NSHD study members who took part in
this latest data collection for their continuing support. They
thank members of the NSHD scientific and data collection
team at the following centers: MRC Unit for LHA; Medical
Research Council (MRC); Clinical Research Facility (CRF)
Manchester; WTCRF at the Western General Hospital in
Edinburgh; WTCRF at University Hospital Birmingham;
WTCRF at University College London Hospital; CRF at the
University Hospital of Wales; CRF and Twin Research Unit at
St Thomas’ Hospital London.
References
[1] A. Wimo and M. Prince, “Alzheimer’s disease international
world Alzheimer report 2010: the global economic impact of
dementia,” in Alzheimer’ Disease International, London, UK,
2010.
[2] L. J. Whalley, F. D. Dick, and G. McNeill, “A life-course
approach to the aetiology of late-onset dementias,” The Lancet
Neurology, vol. 5, no. 1, pp. 87–96, 2006.
Journal of Aging Research
[3] L. J. Whalley, H. C. Fox, I. J. Deary, and J. M. Starr, “Childhood
IQ, smoking, and cognitive change from age 11 to 64 years,”
Addictive Behaviors, vol. 30, no. 1, pp. 77–88, 2005.
[4] B. M. Van Gelder, M. A. R. Tijhuis, S. Kalmijn, S. Giampaoli,
A. Nissinen, and D. Kromhout, “Physical activity in relation to
cognitive decline in elderly men: the FINE study,” Neurology,
vol. 63, no. 12, pp. 2316–2321, 2004.
[5] I. J. Deary, J. Corley, A. J. Gow et al., “Age-associated cognitive
decline,” British Medical Bulletin, vol. 92, no. 1, pp. 135–152,
2009.
[6] Y. Lee, J. H. Back, J. Kim et al., “Systematic review of health
behavioral risks and cognitive health in older adults,” International Psychogeriatrics, vol. 22, no. 2, pp. 174–187, 2010.
[7] T. K. T. Phung, K. Andersen, L. V. Kessing, and G. Waldemar,
“Lifestyle-related risk factors for dementia,” Ugeskrift for
Laeger, vol. 168, no. 40, pp. 3401–3405, 2006.
[8] L. Fratiglioni and H. X. Wang, “Brain reserve hypothesis in
dementia,” Journal of Alzheimer’s Disease, vol. 12, no. 1, pp.
11–22, 2007.
[9] L. Fratiglioni, B. Winblad, and E. von Strauss, “Prevention of
Alzheimer’s disease and dementia. Major findings from the
Kungsholmen Project,” Physiology and Behavior, vol. 92, no.
1-2, pp. 98–104, 2007.
[10] S. Sabia, H. Nabi, M. Kivimaki, M. J. Shipley, M. G. Marmot,
and A. Singh-Manoux, “Health behaviors from early to late
midlife as predictors of cognitive function,” American Journal
of Epidemiology, vol. 170, no. 4, pp. 428–437, 2009.
[11] Y. Lee, J. Kim, and J. H. Back, “The influence of multiple
lifestyle behaviors on cognitive function in older persons
living in the community,” Preventive Medicine, vol. 48, no. 1,
pp. 86–90, 2009.
[12] N. P. Pronk, L. H. Anderson, A. L. Crain et al., “Meeting
recommendations for multiple healthy lifestyle factors: prevalence, clustering, and predictors among adolescent, adult, and
senior health plan members,” American Journal of Preventive
Medicine, vol. 27, pp. 25–33, 2004.
[13] W. Poortinga, “The prevalence and clustering of four major
lifestyle risk factors in an English adult population,” Preventive
Medicine, vol. 44, no. 2, pp. 124–128, 2007.
[14] G. Mishra, K. Ball, J. Arbuckle, and D. Crawford, “Dietary
patterns of Australian adults and their association with socioeconomic status: results from the 1995 National Nutrition
Survey,” European Journal of Clinical Nutrition, vol. 56, no. 7,
pp. 687–693, 2002.
[15] Y. Ben-Shlomo and D. Kuh, “A life course approach to
chronic disease epidemiology: conceptual models, empirical
challenges and interdisciplinary perspectives,” International
Journal of Epidemiology, vol. 31, no. 2, pp. 285–293, 2002.
[16] D. Kuh, Y. Ben-Shlomo, J. Lynch, J. Hallqvist, and C. Power,
“Life course epidemiology,” Journal of Epidemiology and
Community Health, vol. 57, no. 10, pp. 778–783, 2003.
[17] L. J. Launer, “The epidemiologic study of dementia: a lifelong quest?” Neurobiology of Aging, vol. 26, no. 3, pp. 335–340,
2005.
[18] R. C. Petersen, J. E. Parisi, D. W. Dickson et al., “Neuropathologic features of amnestic mild cognitive impairment,”
Archives of Neurology, vol. 63, no. 5, pp. 665–672, 2006.
[19] M. Lamar, S. M. Resnick, and A. B. Zonderman, “Longitudinal
changes in verbal memory in older adults: distinguishing the
effects of age from repeat testing,” Neurology, vol. 60, no. 1, pp.
82–86, 2003.
[20] S. A. Small, Y. Stern, M. Tang, and R. Mayeux, “Selective
decline in memory function among healthy elderly,” Neurology, vol. 52, no. 7, pp. 1392–1396, 1999.
11
[21] J. T. Becker, S. W. Davis, K. M. Hayashi et al., “Threedimensional patterns of hippocampal atrophy in mild cognitive impairment,” Archives of Neurology, vol. 63, no. 1, pp. 97–
101, 2006.
[22] N. C. Fox, W. R. Crum, R. I. Scahill, J. M. Stevens, J. C.
Janssen, and M. N. Rossor, “Imaging of onset and progression
of Alzheimer’s disease with voxel-compression mapping of
serial magnetic resonance images,” The Lancet, vol. 358, no.
9277, pp. 201–205, 2001.
[23] M. Richards, B. Shipley, R. Fuhrer, and M. E. J. Wadsworth,
“Cognitive ability in childhood and cognitive decline in midlife: longitudinal birth cohort study,” British Medical Journal,
vol. 328, no. 7439, pp. 552–554, 2004.
[24] A. Singh-Manoux, M. Kivimaki, M. M. Glymour et al.,
“Timing of onset of cognitive decline: results from Whitehall
II prospective cohort study,” British Medical Journal, vol. 344,
no. 7840, 2012.
[25] L. Fratiglioni, S. Paillard-Borg, and B. Winblad, “An active and
socially integrated lifestyle in late life might protect against
dementia,” The Lancet Neurology, vol. 3, no. 6, pp. 343–353,
2004.
[26] Y. Stern, “Cognitive reserve,” Neuropsychologia, vol. 47, no. 10,
pp. 2015–2028, 2009.
[27] D. J. L. Kuh and C. Cooper, “Physical activity at 36 years:
patterns and childhood predictors in a longitudinal study,”
Journal of Epidemiology and Community Health, vol. 46, no.
2, pp. 114–119, 1992.
[28] V. Mikkilä, L. Räsänen, O. T. Raitakari, P. Pietinen, and J.
Viikari, “Longitudinal changes in diet from childhood into
adulthood with respect to risk of cardiovascular diseases: the
Cardiovascular Risk in Young Finns study,” European Journal
of Clinical Nutrition, vol. 58, no. 7, pp. 1038–1045, 2004.
[29] V. H. Menec, “The relation between everyday activities and
successful aging: a 6-year longitudinal study,” Journals of
Gerontology B, vol. 58, no. 2, pp. S74–S82, 2003.
[30] H. Iwasa, Y. Yoshida, I. Kai, T. Suzuki, H. Kim, and H.
Yoshida, “Leisure activities and cognitive function in elderly
community-dwelling individuals in Japan: a 5-year prospective cohort study,” Journal of Psychosomatic Research, vol. 72,
no. 2, pp. 159–164, 2012.
[31] M. M. Esiri, Z. Nagy, M. Z. Smith, L. Barnetson, A. D. Smith,
and C. Joachim, “Cerebrovascular disease and threshold for
dementia in the early stages of Alzheimer’s disease,” The
Lancet, vol. 354, no. 9182, pp. 919–920, 1999.
[32] M. Wadsworth, D. Kuh, M. Richards, and R. Hardy, “Cohort
profile: the 1946 National Birth Cohort (MRC National
Survey of Health and Development),” International Journal of
Epidemiology, vol. 35, no. 1, pp. 49–54, 2006.
[33] D. Kuh, M. Pierce, J. Adams et al., “Cohort Profile: updating
the cohort profile for the MRC National Survey of Health and
Development: a new clinic-based data collection for ageing
research,” International Journal of Epidemiology, vol. 40, no. 1,
Article ID dyq231, pp. e1–e9, 2011.
[34] H. L. Taylor, D. R. Jacobs, and B. Schucker, “A questionnaire
for the assessment of leisure time physical activities,” Journal
of Chronic Diseases, vol. 31, no. 12, pp. 741–755, 1978.
[35] R. Cooper, G. D. Mishra, and D. Kuh, “Physical activity across
adulthood and physical performance in midlife: findings
from a British birth cohort,” American Journal of Preventive
Medicine, vol. 41, no. 4, pp. 376–384, 2011.
[36] G. M. Price, A. A. Paul, F. B. Key et al., “Measurement of diet
in a large national survey: comparison of computerized and
manual coding of records in household measures,” Journal of
12
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
Journal of Aging Research
Human Nutrition and Dietetics, vol. 8, no. 6, pp. 417–428,
1995.
C. J. Prynne, A. A. Paul, G. D. Mishra, D. C. Greenberg, and M.
E. J. Wadsworth, “Changes in intake of key nutrients over 17
years during adult life of a British birth cohort,” British Journal
of Nutrition, vol. 94, no. 3, pp. 368–376, 2005.
S. Köhler, M. P. J. Van Boxtel, J. Van Os et al., “Depressive
symptoms and cognitive decline in community-dwelling older
adults,” Journal of the American Geriatrics Society, vol. 58, no.
5, pp. 873–879, 2010.
J. M. Starr, I. J. Deary, H. C. Fox, and L. J. Whalley, “Smoking
and cognitive change from age 11 to 66 years: a confirmatory
investigation,” Addictive Behaviors, vol. 32, no. 1, pp. 63–68,
2007.
M. Richards, M. J. Jarvis, N. Thompson, and M. E. J.
Wadsworth, “Cigarette smoking and cognitive decline in
midlife: evidence from a Prospective Birth Cohort study,”
American Journal of Public Health, vol. 93, no. 6, pp. 994–998,
2003.
Office of Population Censuses and Surveys, Classification of
Occupations, HMSO, London, UK, 1970.
D. A. Pigeon, “Tests used in the 1954 and 1957 surveys,” in
The Home and the School (appendix 1), M. Wadsworth, Ed.,
Macgibbon & Kee, London, UK, 1964.
M. Lindelow, R. Hardy, and B. Rodgers, “Development of a
scale to measure symptoms of anxiety and depression in the
general UK population: the psychiatric symptom frequency
scale,” Journal of Epidemiology and Community Health, vol. 51,
no. 5, pp. 549–557, 1997.
D. P. Goldberg and V. F. Hillier, “A scaled version of the
General Health Questionnaire,” Psychological Medicine, vol. 9,
no. 1, pp. 139–145, 1979.
T. A. Salthouse, “The processing-speed theory of adult age
differences in cognition,” Psychological Review, vol. 103, no. 3,
pp. 403–428, 1996.
S. Belleville, N. Rouleau, and N. Caza, “Effect of normal aging
on the manipulation of information in working memory,”
Memory and Cognition, vol. 26, no. 3, pp. 572–583, 1998.
K. J. Anstey, C. Von Sanden, A. Salim, and R. O’Kearney,
“Smoking as a risk factor for dementia and cognitive decline:
a meta-analysis of prospective studies,” American Journal of
Epidemiology, vol. 166, no. 4, pp. 367–378, 2007.
D. T. Nash, “Nutritional and exercise aspects of cognitive
impairment,” Journal of Clinical Lipidology, vol. 1, no. 4, pp.
242–247, 2007.
K. Yaffe, D. Barnes, M. Nevitt, L. Y. Lui, and K. Covinsky,
“A prospective study of physical activity and cognitive decline
in elderly women women who walk,” Archives of Internal
Medicine, vol. 161, no. 14, pp. 1703–1708, 2001.
M. Aickin, “Dealing with change: using the conditional change
model for clinical research,” The Permanente Journal, vol. 13,
no. 2, pp. 80–84, 2009.
M. Richards, R. Hardy, and M. E. J. Wadsworth, “Does active
leisure protect cognition? Evidence from a national birth
cohort,” Social Science and Medicine, vol. 56, no. 4, pp. 785–
792, 2003.
S. Sabia, A. Elbaz, A. Dugravot et al., “Impact of smoking
on cognitive decline in early old age: the Whitehall II cohort
study,” Archives of General Psychiatry, vol. 69, no. 6, pp. 627–
635, 2012.
S. Rovio, I. Kåreholt, E. L. Helkala et al., “Leisure-time physical
activity at midlife and the risk of dementia and Alzheimer’s
disease,” The Lancet Neurology, vol. 4, no. 11, pp. 705–711,
2005.
[54] Y. E. Geda, R. O. Roberts, D. S. Knopman et al., “Physical
exercise, aging, and mild cognitive impairment a populationbased study,” Archives of Neurology, vol. 67, no. 1, pp. 80–86,
2010.
[55] M. T. Sturman, M. C. Morris, C. F. Mendes De Leon,
J. L. Bienias, R. S. Wilson, and D. A. Evans, “Physical
activity, cognitive activity, and cognitive decline in a biracial
community population,” Archives of Neurology, vol. 62, no. 11,
pp. 1750–1754, 2005.
[56] B. Shatenstein, G. Ferland, S. Belleville et al., “Diet quality
and cognition among older adults from the NuAge study,”
Experimental Gerontology, vol. 47, no. 5, pp. 353–360, 2012.
[57] E. Kesse-Guyot, H. Amieva, K. Castetbon et al., “Adherence
to nutritional recommendations and subsequent cognitive
performance: findings from the prospective Supplementation
with Antioxidant Vitamins and Minerals 2 (SU.VI.MAX 2)
study,” American Journal of Clinical Nutrition, vol. 93, no. 1,
pp. 200–210, 2011.
[58] N. Scarmeas, Y. Stern, M. X. Tang, R. Mayeux, and J.
A. Luchsinger, “Mediterranean diet (MeDi) and risk of
Alzheimer’s disease (AD),” Neurology, vol. 66, no. 5, p. 309,
2006.
[59] V. Solfrizzi, V. Frisardi, D. Seripa et al., “Mediterranean diet
in predementia and dementia syndromes,” Current Alzheimer
Research, vol. 8, no. 5, pp. 520–542, 2011.
[60] Y. Gu, J. W. Nieves, Y. Stern, J. A. Luchsinger, and N. Scarmeas,
“Food combination and alzheimer disease risk: a protective
diet,” Archives of Neurology, vol. 67, no. 6, pp. 699–706, 2010.
[61] B. Hanseeuw and A. Ivanoiu, “Performance on the RI-48
cued recall test best predicts conversion to dementia at the
5- and 10-year follow-ups,” Dementia and Geriatric Cognitive
Disorders Extra, vol. 1, no. 1, pp. 258–266, 2011.
[62] E. Aretouli, O. C. Okonkwo, J. Samek, and J. Brandt, “The fate
of the 0.5s: predictors of 2-year outcome in mild cognitive
impairment,” Journal of the International Neuropsychological
Society, vol. 17, no. 2, pp. 277–288, 2011.
[63] M. Hamer and Y. Chida, “Physical activity and risk of neurodegenerative disease: a systematic review of prospective evidence,” Psychological Medicine, vol. 39, no. 1, pp. 3–11, 2009.
[64] S. J. Colcombe, K. I. Erickson, N. Raz et al., “Aerobic
fitness reduces brain tissue loss in aging humans,” Journals of
Gerontology A, vol. 58, no. 2, pp. 176–180, 2003.
[65] J. Prickaerts, G. Koopmans, A. Blokland, and A. Scheepens,
“Learning and adult neurogenesis: survival with or without
proliferation?” Neurobiology of Learning and Memory, vol. 81,
no. 1, pp. 1–11, 2004.
[66] C. W. Cotman and C. Engesser-Cesar, “Exercise enhances and
protects brain function,” Exercise and Sport Sciences Reviews,
vol. 30, no. 2, pp. 75–79, 2002.
[67] S. A. Neeper, F. Gómez-Pinilla, J. Choi, and C. W. Cotman,
“Physical activity increases mRNA for brain-derived neurotrophic factor and nerve growth factor in rat brain,” Brain
Research, vol. 726, no. 1-2, pp. 49–56, 1996.
[68] A. Edman, L. Edenbrandt, J. Freden-Lindqvist, M. Nilsson,
and A. Wallin, “Asymmetric cerebral blood flow in patients
with mild cognitive impairment: possible relationship to
further cognitive deterioration,” Dementia and Geriatric Cognitive Disorders Extra, vol. 1, no. 1, pp. 228–236, 2011.
[69] B. V. Zlokovic, “Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other disorders,” Nature
Reviews Neuroscience, vol. 12, no. 12, pp. X723–X738, 2011.
[70] W. W. Spirduso, “Physical fitness, aging, and psychomotor
speed: a review,” Journals of Gerontology, vol. 35, no. 6, pp.
850–865, 1980.
Journal of Aging Research
[71] M. Woodward, C. Bolton-Smith, and H. Tunstall-Pedoe,
“Deficient health knowledge, diet, and other lifestyles in
smokers: is a multifactorial approach required?” Preventive
Medicine, vol. 23, no. 3, pp. 354–361, 1994.
[72] N. Scarmeas, J. A. Luchsinger, N. Schupf et al., “Physical
activity, diet, and risk of Alzheimer disease,” Journal of the
American Medical Association, vol. 302, no. 6, pp. 627–637,
2009.
13