European Journal of Clinical Nutrition (1997) 51, 387±393
ß 1997 Stockton Press. All rights reserved 0954±3007/97 $12.00
The association between dietary patterns and cardio vascular
disease risk indicators in healthy youngsters: Results covering
®fteen years of longitudinal development
GB Post, HCG Kemper, J Twisk and W van Mechelen
EMGO institute, Faculty of Medicine, Amsterdam Growth and Health Research Group, Vrije Universiteit, Amsterdam, The Netherlands
Objective: To examine longitudinal relationships between nutrition and risk indicators for cardio vascular
diseases (CVD) during adolescence and young adulthood.
Design: A longitudinal study over ®fteen years.
Subjects: 98 females and 84 males, from 13 to 27 years.
Methods: By means of six interviews dietary patterns were determined. Blood samples were analyzed for serum
concentration of total cholesterol (TC), and high-density-lipoprotein (HDL), bloodpressure, body fat and
maximal aerobic power (VO2max) were determined. The longitudinal relations were analyzed with generalized
estimation equations (GEE), a statistical technique in which relations at different time-points are tested
simultaneously.
Results: Compared to Dutch recommendations six out of seven macro nutrients appear to be unfavorable with
respect to CVD. Borderline or high CVD risk values are apparent at 27 y in more than 25% of the subjects with
respect to percentage body fat and serum total cholesterol in both sexes. In males 40% or more show borderline
hypertension. The `univariate' longitudinal analyses showed signi®cantly positive relations: (1) between the
intake of animal protein, saturated fat (SFA), cholesterol (Chol) and TC, and HDL; (2) between total energy
intake (EN) and systolic blood pressure, and VO2max. Signi®cantly negative associations were found: (1)
between EN, poly-unsaturated fat (PUFA) and TC concentrations; (2) between EN and sum of four skinfolds
(SSF).
Conclusions: With increasing age, over a period of 15 y in both sexes the SFA and Chol intake relate
signi®cantly to the development of a negative CVD risk pro®le. The intake of PUFA relates positive to a CVD
risk pro®le. The signi®cantly negative relation between EN intake and body fat (SSF) is partly explained by the
relation between EN and VO2max.
Descriptors: diet composition; cardio vascular risk indicators; young females and males; longitudinal study
Introduction
A diet rich in saturated fat, cholesterol and animal protein is
a major factor leading to elevations of total serum cholesterol, whereas a high intake of poly-unsaturated fat (PUFA)
plays a role in lowering serum cholesterol levels. High
energy intake relates to obesity, and this affects blood
pressure. Therefore proper diets appear to exert a favorable
in¯uence on CVD by modifying established risk factors
and reducing adiposity. (Posner et al, 1993; Shea et al,
1993; Clevidence et al 1992; Anderson et al, 1991; Barr et
al, 1992; Edelstein et al, 1992; Grundy & Denke, 1990).
In epidemiological research the identi®cation of risk
indicators for CVD in adults is well established. More
recently interest has focused on CVD risk factors in
children and young adults, supposing that atherosclerosis
starts early in childhood (Bild et al, 1996; Kelder et al,
1994). Longitudinal studies investigating the relation
between dietary patterns and CVD risk indicators in
young females and males are scarce (Twisk et al, 1996;
Liu et al, 1996; Hunter et al, 1995).
Correspondence: Dr. GB Post, EMGO Institute/Faculty of Medicine/
AGGO, Vrije Universiteit, v.d. Boechorststraat 7, 1081 BT, Amsterdam,
The Netherlands
Received 30 June 1996; revised 1 March 1997; accepted 7 March 1997
The study of potentially deleterious patterns of diet
affecting blood lipids and body composition within individuals requires long-term serial measurements. In the
`Amsterdam Growth and Health Study' longitudinal observations were made over a period of 15 y in approximately
200 male and female subjects, from the teenage period until
the adult age, that is between the ages of 13 and 27 y.
(Kemper, 1995).
In previous analyses a direct relationship could not be
demonstrated in these males and females, between, on the
one hand, total energy intake and the energy contributed by
total fat and, on the other hand, the sum of four skin folds
and the serum cholesterol level. (Post, 1989; Post & Welten,
1995). These ®ndings were explained by the fact that the
level of risk factors for hyper-cholesterolemia and obesity
was relatively low and did not approach adult risk levels yet.
The longitudinal development of CVD risk indicators,
however, will not solely be in¯uenced by nutritional factors
such as the total intake of energy, fat or saturated fat, but
also by biological parameters like cardiopulmonary ®tness
and body weight together with other lifestyle factors like
physical activity behaviour (Twisk et al, 1996).
This present study explores the longitudinal association
between nutritional factors, namely the intake of energy,
animal protein, saturated fat, poly-unsaturated fat and
cholesterol, as independent variables and CVD risk indica-
Amsterdam Growth and Health Study
GB Post et al
388
tors, namely systolic-, diastolic blood pressure, serum
lipids and sum of skin folds, as dependent variables,
during adolescence and young adulthood over a period of
15 y.
Subjects and methods
Study population
The study population reported here consists of the longitudinal sample of the Amsterdam Growth and Health
Study (Kemper, 1985; Kemper, 1995). In 1977 the study
started with 307 pupils of a secondary school in Amsterdam
(The Netherlands). During the adolescent period (1977±
1980) there was a drop-out rate of 24% resulting in four
measurements taken yearly of 131 girls and 102 boys. Five
years later, in 1985, a total of 200 subjects was remeasured.
In 1991, a sixth measurement took place on 182 subjects,
98 females and 84 males. No signi®cant drop-out effects
were demonstrated with respect to the parameters relevant
to the present report (Post & Kemper, 1995).
The multiple longitudinal design of the study is
described in detail by Kemper (1985, 1995).
The present analysis considers the longitudinal data of
the 98 females and 84 males from 13 until 27 y. The quality
controls of the measured variables was done in two ways:
®rstly all measurements were executed by the same trained
testers according the same guidelines over all the years;
secondly the test-retest reproducibility of each measurement is checked by a statistical procedure calculating the
interperiod correlations between the repeated measurements. All reproducibility coef®cients showed suf®cient
values (> 0.60) used in the variables used in this paper. The
details are published elsewhere (Kemper et al, 1985; Post,
1989).
Cardio vascular disease indicators
Risk indicators for CVD used as dependent variables in this
study were the sum of four skin folds (SSF), systolic and
diastolic blood pressure (SBP and DBP), serum total
cholesterol (TC) and serum high-density lipoproteins
(HDL) concentrations and maximal aerobic power.
Between ages 13 and 27 y, six measurements were taken
of anthropometric and other biological characteristics.
Body fatness was de®ned by the SSF, namely by the sum
of the bicipital, tricipital, sub scapular and supra-iliacal
skinfolds (Harpenden skinfold caliper) (Durnin & Ramahan, 1967; Weiner & Lourie, 1969). The SBP and DBP
were measured with a sphygmomanometer and a standard
pressure cuff at the left upper arm with the subject in a
sitting position.
Non-fasting TC and HDL were analyzed enzymatically
(Lipid Laboratory at the Department of Human Nutrition,
Agriculture University, Wageningen, The Netherlands).
Analyses met the criteria of the WHO reference laboratories: TC was determined according to the methods of
Huang et al (1961) and Abell et al (1952), the HDL
cholesterol fraction was analyzed according to the
method of Burstein & Samaille (1960). Cardio respiratory
®tness was determined by measuring maximal aerobic
power (VO2max). It was assessed with a direct method
using a standard running test on a treadmill until exhaustion
and was expressed in ml min71. kg72/3 (Kemper, 1985).
The longitudinal development of several CVD risk
indicators are different for females and males, and in¯uenced by sexual maturation (Berenson et al, 1981). Before
full maturity is reached biological age can be different from
calendar age. Rapidly maturing subjects have higher biological ages than calendar ages, and slowly maturing
subjects have lower biological ages than calendar ages.
During the adolescent period of the study biological age
was determined by measuring skeletal age from X-rays of
the left hand, according to the Tanner-Whitehouse II
method (Tanner et al, 1975). In the statistical analyses,
corrections are made for gender and biological age.
Dietary assessment
Daily food intake data were collected with a detailed crosscheck dietary history interview (Burke, 1947; Beal, 1967)
adapted to the Amsterdam Growth and Health longitudinal
study (Post, 1989). This method provided information
about the usual food consumption pattern preceding the
interview.
The measurement of food intake was consistently executed in the same period of the year, namely in the months
January until May, to avoid seasonal bias, by the same
dietician. Study participants were asked to recall their usual
food consumption pattern during the past month. Frequency, amounts and methods of preparations of the
foods consumed were assessed in household measures or
common portion sizes. A variety of visual aids were used in
depicting portion sizes. Only the food items eaten at least
twice monthly were recorded. Information on foods consumed during regular meals as well as in between meals
(so-called snacks) were collected separately for the ®ve
school or work days and for the two weekend days. A
Dutch food-composition database was used in calculating
energy and mean daily nutrient intakes (NEVO Table,
1991). As far as the concerned nutrients in this study the
percentage of missing values are below 1%. A detailed
description of the dietary assessment methodology and the
reproducibility of the dietary history used in this study was
reported previously. (Kemper, 1995; Post, 1989).
Dietary characteristics as independent variables relevant
for the present analysis included the intake of total energy
(En, in MJ), animal protein (AProt), saturated fat (SFA),
poly unsaturated fat (PUFA), and cholesterol (Chol).
Because a higher energy intake mostly results in a higher
intake of nutrients, the nutritional factors AProt, SFA,
PUFA, and Chol were adjusted for energy intake.
Data analysis
A longitudinal correlation coef®cient was calculated
between nutritional factors and risk indicators for CVD
using the generalized estimating equations (GEE) technique. (Zeger & Liang, 1986). The following statistical
model was used:
P
P
Yit ˆ b0 ‡ b1j Xijt ‡ b2t ‡
time-dependent covariates
P
‡
time-independent covariates
where: Yit ˆ observations of subject i at time t; b0 ˆ
intercept; blj ˆ standardized regression coef®cient of independent variable j; Xijt ˆ independent variable j of subject i
at time t; b2 ˆ regression coef®cient of time.
This model tests the relation between separate risk
indicators for CVD (Yit) and each of the nutritional factors
(Xit) simultaneously at each point in time, resulting in a
standardized regression coef®cient (b), which can be interpreted as a longitudinal correlation coef®cient. This method
corrects for time-dependent covariates, like year of measurement (time) and biological age, and time-independent
covariates like gender. Because the longitudinal character-
Amsterdam Growth and Health Study
GB Post et al
istics of the same subject (Yit) are not independent of each
other, the parameters of the model have to be estimated
adjusting for these intra-subject correlations. With GEE the
relations were calculated between the longitudinal development of the risk indicators for CVD (Yit; TC, HDL, SBP,
DBP, SSF and VO2max) and separate of the dietary
parameters (Xit; EN, AProt, SFA, PUFA, Chol.) over the
entire longitudinal period of 15 y. First of all, the nutritional factors were entered individually into the statistical
model and if signi®cant relations (P 0.05) were found
with more than one of the nutritional factors, a multivariate
GEE was carried out.
Differences between the mean intakes of the subjects
with Dutch recommended daily allowances (RDAs) are
tested with paired t-tests.
Results
The EN intake in females ( 9.5 MJ) showed no signi®cant age trend between 13 and 27 y (see Figure 1). The
males showed during the adolescent period a signi®cant
increase until adult age (from 11.6 MJ to 13.8 MJ,
P 0.001), and thereafter a signi®cant decrease appeared
at adult age (12.7 MJ). Throughout the same age period the
contribution of protein to the energy intake increased from
13±14% in females and in males from 12±13% (both
P 0.001), in particular due to the AProt intake. The
intake of carbohydrates decreased with increasing age,
from about 47±44 energy per cent in females, and from
46 to 45 energy per cent in males (P 0.01). The contribution of total daily fat intake decreased (P 0.01) from
about 42% of total energy to about 39.5%, in females as
well as in males. Finally during this age period the intake of
Chol ¯uctuated between the 26±33 mg/MJ, with a peak
intake at young adult age (females 33 mg/MJ, males
30 mg/MJ).
The CVD risk indicators showed in our population with
increasing age increasing value in females and males for
SSF (see Figure 2), blood pressure (see Figure 3), and TC
and HDL concentration (see Figure 4). The SSF increased
in females about 40% (P 0.001), between the age 13 and
21 y, followed by a signi®cant decrease at age 27 y (Figure
Figure 1 The longitudinal change in total daily energy intake (MJ) (mean and s.d.) for females and males from 13±27 y of age.
Figure 2 The longitudinal change in sum of skinfolds (mm) (mean and s.d.) for females and males from 13±27 y of age.
389
Amsterdam Growth and Health Study
GB Post et al
390
Figure 3 The longitudinal change in systolic bloodpressure (SBP) and diastolic bloodpressure (DBP) in mm Hg (mean and s.d.) for females and males
from 13±27 y of age.
2). Males showed a gradual increase in the SSF from the
age of 15 y until the age of 27 y of about 30% (P 0.001).
In females the SBP remained more or less constant
through adolescence and young adulthood, and the DBP
increased slightly but signi®cantly (P 0.001) (Figure 3).
In males a gradual increasing SBP was found (P 0.001)
with increasing age. DBP increased by about 10%
(P 0.001) from age 21 until age of 27 y.
The serum TC and HDL concentration in females
showed hardly any change between the ages of 13 and
16 y, but over the young adult and the adult period (®fth
and sixth measurement) the TC concentration increased by
almost 20% (P 0.001) (Figure 4). The HDL concentration
increased only during the adult period (P 0.001). Initially,
during adolescence, males TC concentration decreased by
10%, followed by a 10% increase thereafter (both
P 0.001). The HDL concentration decreased between
ages 13 and 21 y (P 0.001), with no further changes at
age 27 y.
The signi®cant relations, resulting from the `univariate'
GEE analyses between the longitudinal changes of the
nutritional factors on the one hand and the longitudinal
changes of CVD indicators on the other hand, are shown in
Table 1.
The GEE analyses showed no interaction with gender,
so the relations were not different for males and females.
The univariate analysis showed that the changes in EN
intake were signi®cantly but negatively related to the
changes in the SSF (P 0.015); positively related to
blood pressure (especially the SBP, P 0.04); negatively
Figure 4 The longitudinal change in serum concentration of total cholesterol (TC) and high density lipoproteins (HDL) (mmol/L) (mean and s.d.) for
females and males from 13±27 y of age.
Amsterdam Growth and Health Study
GB Post et al
Table 1 The signi®cant associations estimated by the GEE `univariate' analyses regarding the longitudinal development of CVD risk indicatorsa in
relation to nutritional characteristicsb, corrected for time, gender and biological age. The Amsterdam Growth and Health Study
EN (MJ)
AProt (% En)
SFA (% En)
c
DBP
SBP
TC
HDL
SSF
VO2max
b
95% CId
P-valuee
b
95% CI
P-value
b
95% CI
P-value
b
95% CI
P-value
b
95% CI
P-value
b
95% CI
P-value
0.075
0.00 to 0.15
0.040
70.096
70.16 to 70.04
0.002
PUFA (% En)
Chol (mg/MJ)
70.076
70.12 to 70.03
0.001
0.107
0.05 to 0.16
0.000
0.080
0.02 to 0.14
0.013
70.067
0.00 to 70.13
0.049
0.073
0.02 to 0.13
0.012
0.083
0.04 to 0.13
0.001
0.089
0.03 to 0.14
0.002
70.073
70.12 to 70.02
0.005
0.105
0.05 to 0.16
0.000
a
TC, serum total cholesterol concentration; HDL, serum high density lipoproteins cholesterol concentration; DBP, diastolic blood pressure; SBP, systolic
blood pressure; TC, serum total cholesterol; SSF, sum of 4 skinfolds; VO2max, maximal aerobic power
EN, total energy intake; AProt, animal protein intake; SFA, saturated fat intake; PUFA, poly-unsaturated fat intake; Chol, cholesterol intake; MJ,
MegaJoules; % En, percentage of energy
c
Correlation coef®cient
d
95% con®dence interval
e
Signi®cant 0.05.
b
related to TC concentration (P 0.012); and positively
related to VO2max (P 0.001).
The changes in intake of AProt (En%) showed a positive
relation to the changes in TC concentration (P 0.012).
The longitudinal changes in intake of SFA (En%) and of
Chol (mg/MJ) were signi®cantly positively related to TC
(respectively P 0.001 and P 0.001) as well as to HDL
concentration (respectively P 0.012 and P 0.013). SFA
intake was also related to the changes in DBP, but the
relationship was negative (P 0.05). The intake of PUFA
(En%) showed over time a negative relation with the TC
concentration (P 0.001). No other signi®cant relations
could be demonstrated.
Based on the signi®cant associations found with the
univariate GEE (Table 1) a multivariate GEE was analyzed
relating the longitudinal changes in the TC and HDL
concentration to the relevant nutritional characteristics,
namely EN, APROT, SFA, PUFA and CHOL (summarized
in Table 2).
The longitudinal changes in Chol intake and SFA intake
remained positively associated with the TC concentration
(P 0.03). The EN and PUFA intake remained inversely
related to TC (respectively P 0.012 and P 0.04). The
relation of AProt with TC disappeared. The association
between the changes in SFA intake and the HDL concentration was positive (P 0.014). For Chol intake the relation with HDL concentration was not signi®cant anymore.
Table 3 summarizes the comparison between the average composition of the diet and the RDAs, in both sexes
over the entire age period. At all ages the differences were
signi®cant (paired t-test, P < 0.000) from the RDAs in an
unfavorable direction. The average intake of total protein,
fat and saturated fat was higher than recommended for a
prudent diet, but lower for the PUFA, total carbohydrate
and poly saccharides intakes. The cholesterol intake was
the only parameter that showed a signi®cant lower and
favorable value with respect to the RDAs. One exception
was found for females at age 21: with a signi®cant higher
value (0.4 mg/MJ, P < 0.000).
Table 4 summarizes the percentages of females and
males with borderline and/or high risk values for CVD at
age 27 y. The results show that less than 5% of the females
have high risk values for SBP and DBP and HDL. In males
this is only shown for SBP and HDL. In 11% of the females
and 12% of the males a borderline DBP is found, and an
additional 12% of the males have a hypertensive DBP. A
borderline SBP was measured in 9% of the females and in
42% of the males (Kemper, 1995).
Table 2 The multivariate signi®cant correlations estimated by GEE regarding the longitudinal development of HDL and TCa concentration in relation to
nutritional factorsb, corrected for time, gender and biological age. The Amsterdam Growth and Health Study
EN (MJ)
c
TC
HDL
a
b
95% CId
P-valuee
b
95% CI
P-valuee
70.096
70.16 to 70.04
0.002
AProt (% En)
SFA (% En)
PUFA (% En)
Chol (mg/MJ)
0.060
0.00 to 0.12
0.030
0.072
0.02 to 0.13
0.014
70.049
70.10 to 0.00
0.042
0.061
0.01 to 0.12
0.033
TC, serum total cholesterol concentration; HDL, serum high density lipoproteins cholesterol concentration
EN, total energy intake; AProt, animal protein intake; SFA, saturated fat intake; PUFA, poly-unsaturated fat intake; Chol, cholesterol intake; MJ,
MegaJoules; % En, percentage of energy
c
Correlation coef®cient
d
95% con®dence interval
e
Signi®cant 0.05.
b
391
Amsterdam Growth and Health Study
GB Post et al
392
Table 3 The range of mean differences in intake of seven macronutrients with the Recommended Daily Allowances
(RDA) given by the Netherlands Nutrition Council (1989), during adolescence and young adulthood in females and
males. Statistical signi®cant differences at all ages with RDAs are indicated
Nutrient
RDA
Range of difference during study
females
Total protein
Total fat
Saturated fat
Poly unsaturated fat
Cholesterol
Total carbohydrate
Poly saccharides
11 en%
35 en%
10 en%
10 en%
33 mg/MJ
55 en%
30 en%
‡1.4
‡4.6
‡5.5
73.5
74.0
78.4
77.1
Table 4 Percentage of subjects above (or below for HDL) borderline
and high risk values for cardio vascular diseases at age 27 y. (Bell et al,
1986; WHO/ISH, 1989; Erkelens, 1991)
Value
Risk factor
Bodyfat (%)
SBP (mm Hg)
DBP (mm Hg)
TC (mmol/l)
HDL (mmol/l)
Borderline
females
males
females
males
females
males
females
males
females
males
25
15
140
140
90
90
5.2
5.2
1.2
1.2
33%
29%
9%
42%
11%
12%
39%
25%
7%
49%
High
30
20
160
160
95
95
6.2
6.2
0.9
0.9
15%
10%
Ð
2%
3%
12%
14%
11%
1%
5%
Discussion
Environmental factors, such as lifestyle, are suggested to be
the driving forces behind the etiology of CVD. Among
these, diet is considered to be highly important. The
western diet, with a relatively high fat intake and low in
complex carbohydrates, is seen as one of the relatively
well-established factors in elevating the risks of CVD
(Posner et al, 1993). In this study, females and males,
showed a decreasing total fat intake, as well as a decreasing
contribution of fat intake to the total energy consumption
with increasing age. Nevertheless, at adult age, this study
population, on average, exceeds the recommended daily
allowances (Netherlands Nutrition Council, 1989). Even
the proportion of subjects, at the age of 27 y, consuming a
high daily fat diet ( 40 en%) was 49% in females and
46% in males (Post and Welten, 1995).
Investigating the long term effects of a dietary pattern
with respect to the development of CVD risk indicators, one
has to make an estimate of the qualitative aspects of the daily
diet over the 15 y period. This was done by comparison with
Dutch RDAs (Netherlands Nutrition Council, 1989). In
relation to CVD risk indicators it could be shown that in
six out of seven selected macro nutrients (see Table 3) the
intakes did not meet the recommendations in both sexes. A
prudent diet, as recommended by the Dutch Nutrition
Council, includes carbohydrates of > 55 En% as well as
30 EN% polysaccharides, because of the inverse relation
with the proportion of energy derived from fat.
Overall it can be stated that the longitudinal dietary
pattern in this population is potentially unhealthy with
respect to the development of CVD, because most of the
macronutrients do not meet the recommended daily allowances in the Netherlands (1989). The literature de®nes
health indicators and risk factors for CVD. (Bell et al,
1986; WHO/ISH, 1989; Erkelens, 1991).
to
to
to
to
to
to
to
‡3.3
‡6.7
‡8.8
74.2
‡0.4
711.2
78.5
males
‡0.5 to ‡2.6
‡4.1 to ‡7.9
‡6.6 to ‡8.5
73.2 to 74.0
73.2 to 77.0
79.0 to 710.6
77.0 to 77.8
Signi®cance in all
years (both sexes)
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
High TC is seen as a major risk factor in relation to
CVD. The cut-off point for ideal serum TC concentration in
the Netherlands is recommended at 5.0 mmol/l. (Erkelens, 1991). The mean TC concentration of males was equal
to this level, whereas females, at the age of 27 y, showed
higher concentrations (see Figure 4). If the `moderate' risk
cut-off point of 5.2 mmol/l is considered, then 25% of the
adult males, and 39% of the females were above this point,
and another 11% of the males and 14% of the females were
above the high risk cut-off point of 6.2 mmol/l. Very low
HDL concentration are seen in 1% of the females and 5%
of the males, but if borderline low values are included, 7%
and 49% respectively are at risk (Table 4). Although
cholesterol intake was the only nutrient that showed favorable lower intakes than RDAs this did not ®nd expression
in lower percentages of subjects with hypercholesteremia
compared to the other CVD risk factors.
Body fat can be estimated from the sum of skin folds. If
one accepts a percentage of body fat of 30% in females and
20% in males as high risk values for overweight, (see Table
4), then 15% of the female subjects and 10% of the male
subjects in our longitudinal study should be regarded as
such at age 27 y. At this age 48% of the females and 39%
of the males have body fat percentages greater than 25%
and 15% respectively. (Kemper, 1995). Although at age
27 y the mean values of our population did not exceed the
high risk values, borderline values for CVD risk indicators
can be seen rather frequently.
In this study the longitudinal relations between nutritional parameters as independent variables and biological
risk indicators for CVD as dependent variables were
analyzed with generalized estimating equations. The
advantage of using GEE is that all longitudinal data are
used and the relation can be analyzed under correction of
both time-dependent, that is year of measurement, biological age, and time-independent, namely gender, co-variates.
(Twisk et al, 1994).
The observed negative association between the TC
concentration and the PUFA intake (declining CVD risk
indicators with age), as well as the positive relation with the
intake of SFA and Chol (increasing CVD risk indicators
with age), support the literature on these aspects. (Hegsted
et al, 1965; Lewis et al, 1981; Weisweiler et al, 1985;
Rosenbergh & Schaefer, 1988; Li-Ching et al, 1994).
Miller et al (1990) found no relationship between EN
intake and adiposity, but demonstrated a higher fat contribution (En%) in obese adults compared to lean subjects,
as well as a lower score on exercise activities. In our study
an inverse relation was found of the SSF to EN intake
(P 0.005) (see Table 1), and a positive association
between EN intake and VO2max (P 0.000). The negative
relation of the SSF and the EN intake can possibly be
Amsterdam Growth and Health Study
GB Post et al
explained by the positive relation of VO2max with EN,
suggesting that subjects with a high VO2max are more
physically active, expend more energy and are as a result
leaner. Therefore the VO2max and EN were taken into a
multivariate GEE for the dependent variable SSF. Although
the negative association between SSF and EN decreased
(from 70.073 to 70.054, 95% CI 70.003 to 70.105), it
remained signi®cant (P 0.036). So, VO2max can only
partly explain this relation.
Conclusions
(1) In comparison with Netherlands Nutrition Council
recommendations, the intake of six out of seven
macro nutrients appeared to be unfavorable with
respect to the development of CVD risk indicators
with increasing age, in both sexes.
(2) At age 27 y, in this study, the prevalence of borderline
CVD risk values is more than 25% for body fat, and
serum total cholesterol in both sexes, but even higher in
males for systolic-, diastolic blood pressure and serum
high density lipoproteins.
(3) In this study a dietary pattern with a relatively high
saturated fat and high cholesterol intake, between ages
13 and 27 y in both sexes, showed to be signi®cantly
and positively related to serum total cholesterol.
(4) The longitudinal analysis revealed that the intake of
poly-unsaturated fat was negatively related to serum
total cholesterol
(5) A dietary pattern with a relative high energy intake in
both sexes was related in this study to a more favorable
coronary risk pro®le, namely a lower amount of body
fat, a higher physical ®tness, a lower diastolic blood
pressure and a lower serum total cholesterol.
References
Abell LL, Levy BB & Kendall FE (1952): Simpli®ed method for the
estimation of total cholesterol in serum and demonstration of its
speci®city. J. Biol. Chem 195, 357±366.
Anderson KM, Wilson PW, Odell PM & Kannel WB (1991): An updated
coronary risk pro®le. A statement for health professionals. Circulation
83, 356±362.
Barr SL, Ramakrishnan R, Johnson C, Holleran S, Dell RP & Ginsberg HN
(1992): Reducing total dietary fat without reducing saturated fatty acids
does not signi®cantly lower total plasma cholesterol concentrations in
normal males. Am. J. Clin. Nutr. 55, 675±681.
Beal VA (1967): The nutritional history in longitudinal research. J. Am.
Diet. Ass. 51, 425±432.
Bell RD, Macek M, Rutenfranz J & Saris WHM (1986): Health indicators
and risk factors of cardio vascular diseases during childhood and
adolescence. In: Children and Exercise XII, Vol 17 eds. Rutenfranz J,
Mocellin R, Klimt F, Int. Series Sports Science, Human Kinetics,
Champaign IL.
Berenson GS, Srinivasan SR, Cresanta JL, Foster TA, Webber LS (1981):
Dynamic changes of serum lipoproteins in children during adolescence
and sexual maturation. Am. J. Epidemiol. 113, 157±170.
Bild DE, Jacobs DR jr, Liu K, Williams OD, Hilner JE, Perkins LL,
Marcovina SM, Hulley SB. (1996): Seven-year trend in plasma lowdensity-lipoprotein-cholesterol in young adults: the CARDIA Study.
Ann. Epidemiol. 6, 235±245.
Burke BS (1947): The dietary history as a tool in research. J. Am. Diet.
Ass. 23, 1041±1046.
Burstein M & Samaille J (1960): Sur un dosage rapide du cholesterol lie
aux alpha-et betalipoproteinen du serum. Clin. Chim. Acta. 5, 609±611.
Clevidence BA, Judd JT, Schatzkin A, Muesing RA, Campbell WS, Brown
CC, Taylor PR. (1992): Plasma lipid and lipoprotein concentrations of
men consuming a low-fat, high-®ber diet. Am. J. Clin. Nutr. 55, 689±
694.
Durnin JVGA & Rahaman MM (1967): The assessment of the amount of
fat in the human body from measurements of skinfold thickness. Brit. J.
Nutr. 21, 681±689.
Edelstein SL, Barrett-Connor EL, Wingard DL & Cohn BA (1992): Increased meal frequency associated with decreased cholesterol concentrations;
Rancho Bernardo CA, 1984±1987. Am. J. Clin. Nutr. 55, 664±669.
Erkelens DW (1991): Herziening concensus cholesterol. Ned. T. Geneesk.
135±49, 2337±2340. Dutch.
Grundy SM & Denke MA (1990): Dietary in¯uences on serum lipids and
lipoproteins. J. Lipid Res. 31, 1149±1172.
Hegsted DM, McGandy RB, Myers ML & Stare FJ (1965): Quantitative
effects of dietary fat on serum cholesterol in man. Am. J. Clin. Nutr. 17,
281±295.
Huang TC, Chen CP, Wegler V & Raftery A (1961): A stable reagent for
the Liberman-Buchard reaction. Anal. Chem. 33, 1405±1407.
Hunter SM, Bao W & Berenson GS (1995): Understanding the development of behavior risk factors for cardiovascular disease in youth:
the Bogalusa Heart Study. Am. J. Med. Sci. 310, Suppl 1, S114±
S118.
Kelder SH, Perry CL, Klepp KI & Lytle LL (1994): Longitudinal tracking
of adolescent smoking, physical activity, and food choice behaviours.
Am. J. Pub. Health 84, 1121±1126.
Kemper HCG (ed) (1985): Growth Health and Fitness of TeenagersÐ
Longitudinal research in International perspective. Medicine and Sport
Science, Vol 20, Karger.
Kemper HCG (ed) (1995): The Amsterdam Growth StudyÐA longitudinal
analysis of health, ®tness and lifestyle. HKP Sport Science Monograph
Series, Vol 6.
Lewis B, Hammett F, Katan M, Kay RM, Merkx I, Nobels A, Miller NE,
Swan AV (1981): Towards an improved lipid-lowering diet: additive
effects of changes in nutrient intake. Lancet 2, 1310±1313.
Li-Ching L, Shieh MJ, Posner BM, Ordovas JM, Dwyer JT, Lichtenstein
AH, Cupples LA, Dallal GE, Wilson PW, Schaefer EJ. (1994):
Relationship between dietary intake, lipoproteins and apoliproteins in
Taipei and Framingham. Am. J. Clin. Nutr. 60; 765±774.
Lie K, Ruth KJ, Flack JM, Jones-Webb R, Bunke G, Savage PJ, Hulley
SB. (1996): Blood pressure in young blacks and whites: relevance of
obesity and lifestyle factors in determining differences. The CARDIA
Study. Coronary artery risk development in young adults. Circulation
93, 60±66.
Miller MC, Lindeman, AK, Wallace J & Niederpruem M (1990): Diet
composition, energy intake, and exercise in relation to body fat in men
and women. Am. J. Clin. Nutr. 52; 426±430.
Netherlands Nutrition Council (1989): Nederlandse Voedingsnormen.
Voedingsraad, Den Haag. (Dutch).
NEVO Table (1991): De Nederlandse voedingsmiddelen tabel. (Dutch).
Posner BM, Cupples LA, Franz MM & Gagnon DR (1993): Diet and heart
disease risk factors in adult American men and women. Int. J.
Epidemiol. 22, 1014±1025.
Post GB & Kemper HCG (1995): Procedure and subjects. In: The
Amsterdam Growth Study, ed. Kemper HCG. HKP Sport Science
Monograph Series, Vol 6, pp 17±27.
Post GB & Welten DC (1995): The development of nutritional intake during
15 years of follow-up. In: The Amsterdam Growth Study, ed. Kemper
HCG, HKP Sport Science Monograph Series, Vol 6, pp 108±134.
Post GB (1989): Nutrition in adolescenceÐa longitudinal study in dietary
patterns from teenager to adult. PhD Thesis. SO 16, De Vrieseborch,
Haarlem, The Netherlands.
Rosenberg IH & Schaefer EJ (1988): Dietary saturated fatty acids and
blood cholesterol. N. Engl. J. Med. 318, 1270±1271.
Shea S, Melnik TA, Stein AD, Zansky SM, Maylahn C & Basch CE
(1993): Age, sex, educational attainment, and race/ethnicity in relation
to consumption of speci®c foods contributing to the atherogenic
potential of diet. Prev. Med. 22, 203±218.
Tanner JM, Whitehouse RH, Marshall WA, Healy MJR & Goldstein H
(1975): Assessment of Skeletal Maturity and Prediction of Adult Height
(TW2 method). London, Academic Press.
Twisk JWR, Kemper HCG & Mellenbergh GJ (1994): Mathematical and
analytical aspects of tracking. Epidemiol. Rev. 2, 165±183.
Twisk JWR, Kemper HCG, Mellenbergh GJ, van Mechelen W & Post GB
(1996): Relation between the longitudinal development of lipoprotein
levels and lifestyle parameters during adolescence and young adulthood. Ann. Epidemiol. 6, 246±256.
Weiner JS & Lourie JA (1969): Human Biology, a Guide to Field Methods.
IBP Handbook, number 9, Oxford, Blackwell.
Weisweiler P, Janetscheck P & Schwandt P (1985): In¯uence of poly
unsaturated fats and fat restriction on serum lipoproteins in humans.
Metabolism 34, 83±87.
WHO/ISH (1989): 1989 guidelines for the management of mild hypertension: Memorandum from WHO/ISH meeting bulletin of the World
Health Organization, 67±5, 493±498.
Zeger SL & Liang KY (1986): Longitudinal data analysis for discrete and
continuous outcomes. Biometrics 42, 121±130.
393