American Journal of Epidemiology
Copyright © 1996 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 143, No. 6
Printed in U.S.A.
Amount and Intensity of Physical Activity, Physical Fitness, and Serum
Lipids in Men
Jaume Marrugat,1 Roberto Elosua,1 Maria-Isabel Covas,2 Luis Molina,3 Juan Rubies-Prat,4 and
the MARATHOM Investigators
Physical activity improves one's lipid profile and increases physical fitness. The present study was aimed
at determining the association of amount and intensity of leisure time physical activity with serum lipid profile
and physical fitness. A total of 537 healthy men aged 20-60 years were recruited in a quota sampling frame
for measurement of physical activity energy expenditure at two different levels. The Minnesota Leisure Time
Physical Activity Questionnaire was administered. Serum lipid and lipoprotein levels were measured, and all
participants were given an exercise test. Physical activities with an intensity greater than 7 kcal/minute were
significantly associated (p < 0.01) with a higher level of high density lipoprotein (HDL) cholesterol and a lower
atherogenic index (total cholesterol: HDL cholesterol). Independently of other confounding variables, each
average 100 kcal/day expended in leisure time physical activity with an intensity greater than 7 kcal/minute
during the previous year was associated with an increase of 2.09 mg/dl (0.054 mmol/liter) in HDL cholesterol
and a decrease of 0.23 in atherogenic index. However, only physical activity with an intensity greater than 9
kcal/minute was associated with decreases in total cholesterol, non-HDL cholesterol, and log(triglycerides).
Better physical fitness was associated with physical activities of intensities above 5 kcal/minute. There is a
threshold in the intensity of exercise associated with serum lipid profile (7 kcal/minute) and physical fitness
(5 kcal/minute). Above the former threshold, the relation between amount of physical activity and lipid levels
is linear for total cholesterol, HDL cholesterol, non-HDL cholesterol, and atherogenic index and is logarithmic
for triglycerides. Am J Epidemiol 1996; 143:562-9.
alcohol drinking; body mass index; cholesterol; exercise; lipoproteins, HDL cholesterol; physical fitness;
smoking; triglycerides
Physical activity and physical fitness have been
identified as protective factors against the occurrence
(1-3) and progression (4) of coronary heart disease
and against premature mortality (5). Such associations,
among other factors, have been related to improvement in the lipid profile (6).
Health promotion programs (7), medical journal articles (8-10), and the media strongly advise the general population to engage in physical activity. Al-
though agreement on recommendations to improve
physical fitness has been reached (11), the amount and
intensity of physical activity required to achieve a
low-risk serum lipid profile have not yet been established (12).
The MARATHOM (Medida de la Actividad fisica y
su Relacion Ambiental con Todos los lipidos en el
Hombre) Study is aimed at determining the association
of amount and intensity of physical activity with serum lipid profile and physical fitness in healthy men
aged 20-60 years.
Received for publication January 3 1 , 1995, and in final form
January 19, 1996.
Abbreviations: BMI, body mass index; EEPA, energy expenditure
in physical activity; HDL, high density lipoprotein; MARATHOM,
Medida de la Actividad fisica y su Relacion Ambiental con Todos los
lipidos en el Hombre.
1
Departament d'Epidemiologia i Salut Publica, Institut Municipal
d'lnvestigacio Medica, Barcelona, Spain.
2
Laboratori de Referenda de Catalunya, Hospital del Mar, Barcelona, Spain.
3
Servei de Cardiologia, Hospital del Mar, Barcelona, Universitat
Autdnoma de Barcelona, Barcelona, Spain.
4
Departament de Medicina, Hospital del Mar, Barcelona, Universitat Autonoma de Barcelona, Barcelona, Spain.
Reprint requests to Dr. Jaume Marrugat, Departament d'Epidemiologia i Salut Publica, Institut Municipal d'lnvestigacio Medica,
Carrer Doctor Aiguader, 80 08003 Barcelona, Spain.
MATERIALS AND METHODS
To balance the representation of all levels of physical activity, we chose a cross-sectional study design
with a quota sampling frame.
Subjects
Sampling was stratified according to two age groups
(20-40 years and 41-60 years) and two groups of
energy expenditure in physical activity (EEPA): active
(EEPA <300 kcal/day) and very active (EEPA >300
562
Physical Activity, Fitness, and Lipids
kcal/day). This cutoff point was selected because a
physical activity energy expenditure greater than
2,000 kcal/week (approximately 300 kcal/day) has
been previously associated with lower coronary heart
disease mortality (13).
Subjects with a history of neoplasia, cardiovascular
disease, hypertension, or diabetes mellitus were excluded, as well as those with an abnormal basal electrocardiogram, an abnormal effort electrocardiogram,
a body mass index (BMI) (weight (kg)/height (m)2)
over 35, alcohol consumption greater than 80 g/day,
illicit drug abuse, or long-term medication use.
A total of 537 participants were recruited: 278 in the
age group 20-40 years (136 of whom had a total
EEPA of 300 kcal/day or less) and 259 in the age
group 41-60 years (149 of whom had a total EEPA
greater than 300 kcal/day). These four strata were
filled by the quota sampling method in a consecutive
manner, as follows. The "very active" subjects were
recruited among marathon runners living in Barcelona
who had participated in the 1990 Barcelona Marathon
(response rate = 45 percent). The "active" participants
were consecutively recruited among healthy bank employees who regularly visited a preventive health center in Barcelona for cardiovascular disease screening.
The screening included a physical examination, a
chest roentgenogram, a lipid profile and other routine
laboratory measurements, a basal electrocardiogram,
and an exercise test.
Instructions to participants
Participants were asked to fast for 10-14 hours
before blood sampling. Blood samples were obtained
before 9:00 a.m. No special recommendations as to
physical activity or alcohol consumption within the 24
hours prior to blood sampling were given; however,
for subjects who said they had exercised within the
previous 12 hours, sampling was postponed for 24
hours to avoid the acute effect of physical activity on
lipid levels.
Sample size
An a risk of 0.05 and a statistical power of 90
percent in a multiple linear regression model with six
variables were accepted. A squared correlation coefficient (r 2 ) of 0.03 when adjusting for four independent variables could be detected as statistically significant with the 537 study participants (SOLO; BMDP
Statistical Software, Los Angeles, California).
Physical activity assessment
For assessment of the amount and intensity of physical activity performed during the previous year, the
Am J Epidemiol
Vol. 143, No. 6, 1996
563
Minnesota Leisure Time Physical Activity Questionnaire was used. This questionnaire has been extensively described elsewhere (14) and has been validated
for use among Spanish men (15). Briefly, the questionnaire is administered by a trained interviewer, and
out of a list of 64 suggested physical activities, participants are asked to mark those they have undertaken
during the past year. The interviewer spends an average of 20 minutes per individual collecting detailed
information on the marked activities. Each physical
activity has an intensity code with values ranging from
2 to 12. The code is based on the ratio between the
body's metabolic rate during the physical activity and
the basal metabolic rate. It is assumed that a 1-unit
change in the intensity code is approximately equivalent to 1 kcal/minute. Total daily average energy expenditure in leisure time physical activity (total EEPA)
during the past year is calculated as described elsewhere (15).
In the present study, we defined six categories of
physical activity according to their intensity: 2-3 kcal/
minute (e.g., bowling or sailing); 3.5-5 kcal/minute
(walking, golf); 5.5-7 kcal/minute (dancing, swimming); 7.5-9 kcal/minute (climbing stairs, tennis);
9.5-11 kcal/minute (squash, handball); and 11.5-12
kcal/minute (running, soccer). Thus, for each particular
subject, total EEPA = EEPA2_3 + EEPA35_5 +
EEPA55_7 + EEPA75_9 + EEPAg^n + EEPA n 5_12,
where EEPA2_3 represents daily average energy expenditure in leisure time physical activity with an intensity of
2-3 kcal/minute during the past year, and so on.
Laboratory methods
Blood samples were taken in the morning after a
fasting period of 10-14 hours on a Tuesday, Wednesday, Thursday, or Friday. Blood was drawn without
stasis after a period of 20 minutes' rest in a supine
position. Sera were frozen at —40°C until analyzed.
Total cholesterol levels were determined enzymatically (the CHOD-PAP method; Merck, Darmstadt,
Germany). HDL cholesterol was measured after precipitation of apoprotein B-containing lipoproteins with
polyethylene glycol (Immuno, Vienna, Austria). Triglyceride levels were also measured enzymatically
(Merck). The methods were standardized with the
World Health Organization lipid program and the
Monitrol quality control program (Baxter Diagnostics,
Dtidingen, Switzerland). Interassay coefficients of
variation were 2.57 percent, 4.59 percent, and 2.90
percent for total cholesterol, HDL cholesterol, and
triglycerides, respectively. Non-HDL cholesterol was
defined as total cholesterol minus HDL cholesterol
(16) and the atherogenic index as the ratio between
total cholesterol and HDL cholesterol.
564
Marrugat et al.
Exercise test
Statistical analysis
All of the participants underwent a maximum effort
treadmill test according to Bruce's protocol (17). Resting heart rate and blood pressure were recorded and a
12-lead electrocardiogram was taken while the participant lay in a supine position.
The exercise test was maximum effort, and the
indications for stopping were those recommended by
the American College of Sports Medicine (18). Moreover, the Bruce protocol was ended after a maximum
of 18 minutes of effort. Seventy subjects (four active
and 66 very active) achieved the 18-minute limit of the
exercise test, eight of them without reaching their
maximum theoretical heart rate (220 minus age in
years). On the other hand, 73 participants stopped
because of fatigue without reaching their maximum
theoretical heart rate (57 in the active group and 16 in
the very active group); in 12 cases,, the exercise test
was stopped because the subject showed a hypertensive reaction, and in five cases because ST segment
depression exceeded 1 mm on the electrocardiogram.
The rest of the subjects completed the exercise test
after reaching their maximum theoretical heart rate.
Time needed to reach the maximum theoretical
heart rate during the effort test was recorded for calculation of maximal oxygen uptake (VO2max), used
as a fitness indicator (19). For participants who were
not able to reach their maximum theoretical heart rate,
the duration of the effort test was recorded instead.
The chi-squared test was used to compare proportions between groups. Student's / test was used to
compare mean values of continuous variables between
two groups. Kruskal-Wallis and Mann-Whitney U
tests were used instead when the data distribution
departed from normality. Multiple analysis of variance
was used to compare mean values of continuous variables adjusted for age.
The nonparametric Spearman's correlation test was
employed to assess the association between two continuous variables.
Prior to assessment of the association between
EEPA (adjusting for age, alcohol consumption, smoking, and BMI) and the lipid variables, the following
assumptions were tested on residuals for linear regression: linearity, equality of variance, independence of
error, and normality. A plot of the residuals was also
examined to detect and investigate outliers.
Diet questionnaire
A self-administered 3-day recall questionnaire of
food intake, including drinks, was given to all participants, who received specific instructions and recommendations on how to describe the type, dressing,
proportion, and quantity of each dish. When subjects
were unable to specify the exact amount consumed in
grams or liters, they were allowed to give a rough
description such as "a full plate" or "a little piece."
Reported information on food intake was converted
into nutritional data using The Diet Analysis Nutritionist IV software (N Squared Computing, San
Bruno, California). Intake of polyunsaturated, monounsaturated, and saturated fatty acids and caloric intake were assessed. Daily average alcohol consumption in grams was estimated from responses to a
specific questionnaire that referred to consumption
during the previous week (including the weekend).
Other measurements
A questionnaire on smoking habits was administered. Family history of coronary heart disease before
the age of 65 years was also recorded.
RESULTS
The characteristics of the entire sample and of persons in each of the physical activity groups are shown
in table 1. Age-adjusted serum lipid and lipoprotein
concentrations are presented by physical activity
group in table 2. Steep decreases were observed in
total cholesterol, non-HDL cholesterol, triglycerides,
and the atherogenic index from the active group to the
very active group, as well as an increase in HDL
cholesterol. All lipid variables showed statistically significant differences between the very active group and
the active group.
Spearman correlations between demographic factors, dietary and electrocardiographic variables, physical activity levels, and the lipid profile are shown in
table 3.
The association between physical activity (for each
expended 100 kcal/day) and serum lipids and lipoproteins, adjusted for age, alcohol consumption, and
smoking, was assessed using multiple linear regression (table 4). Serum triglyceride data required logarithmic transformation to fulfil the regression assumptions.
EEPA2_3, EEPA 35 _ 5 , and EEPA 55 _ 7 were not
significantly associated with serum lipid levels.
EEPA75_9, EEPA 95 _ U , and EEPA 115 _ 12 were directly associated with HDL cholesterol and inversely
associated with atherogenic index. When the latter
EEPAs were grouped together (EEPA7 5 _ 12 ), the coefficients were 2.09 (standard error = 0.23, p < 0.01)
for HDL cholesterol and —0.227 (standard error =
0.028, p < 0.01) for atherogenic index.
Am J Epidemiol
Vol. 143, No. 6, 1996
Physical Activity, Fitness, and Lipids
Characteristics of the MARATHOM* sample population, by physical activity group, Barcelona, Spain, 1991-1993
TABLE 1.
Active
(n = 285)
EEPA* (kcal/day)
EEPA^
EEPA
3W>
EEPA
S5-7
EEP
V*-9
EEPA
565
9S-11
E E R Wi2
Total EEPA
Age (years)
Body mass indext
Heart rate (beats/minute)
Systolic blood pressure
(mmHg)
Diastolic blood pressure
(mmHg)
VO2max* (ml/kg/minute)
Current smoker (%)
Family history of coronary
heart disease (%)
Alcohol consumption (g/day)
Polyunsaturated fat:
saturated fatt
Very active
All participants
(n=252)
Mean
SD*
Mean
SD
51.24
39.16
33.69
11.84
3.41
26.23
166.32
41.95
26.18
73.02
53.00
47.64
44.99
24.94
15.24
73.03
76.60
9.15
3.16
12.33
91.28
58.22
92.60
48.46
13.83
326.85
588.29
39.51
23.61
65.00
115.28
71.90
108.77
90.93
68.41
279.40
323.67
8.88
2.30
11.49
129.60
15.20
126.11
77.27
40.14
11.14
7.33
76.85
48.04
43.51
(n=537)
Mean
SD
69.87
48.03
61.10
28.88
8.26
166.12
362.67
40.82
24.99
69.28
89.83
60.85
86.25
67.13
48.21
248.24
310.10
9.10
3.08
12.59
14.88
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
127.89
15.13
10.05
8.33
NS
<0.05
77.06
44.06
10.61
8.78
15.61
13.68
P
value
<0.05
20.29
30.35
22.57
22.71
14.30
17.78
NS
<0.05
18.65
20.91
2.06
0.59
1.97
0.60
NS
2.01
0.60
16.76
* MARATHOM, Medida de la Actividad fi'sica y su Relacidn Ambiental con Todos los li'pidos en el Hombre; SD, standard deviation; EEPA,
daily energy expenditure in leisure time physical activity of different intensities: 2-3 kcal/minute, 3.5-5 kcal/minute, 5.5-7 kcal/minute, 7.5-9
kcal/minute, 9.5-11 kcal/minute, and 11.5-12 kcal/minute; VO2max, maximal oxygen uptake (indirect circulation).
t Weight (kg)/height (m)*.
t Dietary polyunsaturated fat:saturated fat ratio.
EEPA 9 5_ n and EEPA U 5_12 were inversely associated with total cholesterol, non-HDL cholesterol, and
log(triglycerides), whereas the other EEPAs were not.
When EEPAs greater than or equal to 9.5 kcal/minute
were grouped together (EEPA95_12), the coefficients
were -2.35 (standard error = 0.77, p < 0.01) for
total cholesterol, —4.40 (standard error = 0.80, p <
0.01) for non-HDL cholesterol, and -0.028 (standard
error = 0.004, p < 0.01) for log(triglycerides).
On the other hand, when total EEPA, regardless of
intensity, was taken into account, coefficients were
-2.44 (standard error = 0.68, p < 0.01) for total
cholesterol, 1.70 (standard error = 0.22, p < 0.01) for
HDL cholesterol, -4.11 (standard error = 0.71, p <
0.01) for non-HDL cholesterol, -0.019 (standard error = 0.002, p < 0.01) for log(triglycerides), and
-0.174 (standard error = 0.024, p < 0.01) for the
atherogenic index.
Dietary saturated and polyunsaturated fatty acids,
both as absolute values and as the ratio of the two,
failed to improve any of the previous models. The
inclusion of BMI in the models did not significantly
TABLE 2. Age-adjusted serum lipid and lipoprotein concentrations (mg/dl) in different physical activity groups, Barcelona,
Spain, 1991-1993*
Active
(n = 285)
Total cholesterol
HDLj cholesterol
Non-HDL cholesterol*
Triglycerides
Atherogenic index§
Very active
{n = 252)
Mean
SDt
Mean
SD
P
value
216.98
45.36
171.61
124.22
5.10
41.67
11.48
42.89
78.34
1.70
201.92
53.32
148.64
85.17
4.00
35.97
12.41
38.20
40.45
1.24
<0.01
<0.01
<0.01
<0.01
<0.01
All participants
(n = 537)
Mean
209.51
49.32
160.18
104.74
4.55
SD
39.92
12.53
42.59
65.68
1.60
* To convert values for total cholesterol, HDL cholesterol, and non-HDL cholesterol to mmol/liter, multiply by 0.02586. To convert values
for triglycerides to mmol/liter, multiply by 0.01129.
t SD, standard deviation; HDL, high density lipoprotein.
t Total cholesterol minus HDL cholesterol.
§ Total cholesterol/HDL cholesterol (no units).
Am J Epidemiol
Vol. 143, No. 6, 1996
566
Marrugat et al.
change the coefficients of the other variables (results
not shown; data available upon request).
Results from the adjusted model for V0 2 max are
also shown in table 4. EEPA 55 _ 7 , EEPA 75 _ 9 , and
EEPA 115 _ 12 were independently associated with a
longer V0 2 max, and EEPA2_3 was associated with a
shorter V0 2 max.
Results of further adjustment of serum lipid models
for V0 2 max are presented in table 5. V0 2 max was
significantly and directly associated with HDL cholesterol and inversely associated with log(triglycerides)
and atherogenic index; no association with total or
non-HDL cholesterol was observed. Inclusion of this
variable in the models did not substantially change the
explained variability of the dependent variables or the
magnitude of the coefficients for the EEPA categories.
i- O
I
ic
•o
a
of
•D
Q.
r^. o r^ co to
CO CO » - T - O
d ci ci d d
DISCUSSION
i- d d d 6 d d
I
!
I I I I
1
I
q o> •? <q c\i T^ c\i
rddddddd
I
•a
I
OCMMOCOOOO)
i
CO CO CO Tf; - ^ CO y-m
I
a
re
I
if
^
i-dddddddd
m
I
oit-^fococomtnco
cucqqcMC\jc\iq'-co
ddddddddd
<
I
si
I
dddddddddd
i l l
I II i
I
a
£
i-
5
I
I
I
III I
i l l
I
i-ddd ddddddddddd
I I I
d dd d
a
CD
IIII
i- d d ddddddddddd
£
I
18 1
I
q qnqqqoqqqqq
d ddddddddddd
1
0)
O)
I III
CO nr-ifNMflO'-Ol^CO
LU
LU
i
ddddddddddd
i l l
HI
I
I
I I
II
I
qqq
ddddddddddd
I
I
I
I
I
O>
5.S.
Ǥ
2i
<a
P CD
= E S
ddddddddddd
i- d d d d d
I
I III
<i>
I
I
Physical activity has been described as "an agent
with lipid-lowering, antihypertensive, positive inotropic, negative chronotropic, vasodilating, diuretic, anorexigenic, weight-reducing, cathartic, hypoglycemic,
tranquilizing, hypnotic and antidepressive qualities"
(8).
Population studies addressing the relations between
both cardiovascular mortality and lipid profile and
physical activity have shown a poor correlation because of the scarce representation of very active people (20-22). Accordingly, we preferred a quota sampling alternative in order to obtain a wide range of
physical activity levels. The hypothesis under consideration was that when the associations between
amount and intensity of physical activity and serum
lipids were examined, a dose-response relation would
be found (23). For more precise assessment of the
effect of physical activity intensity, EEPAs were
grouped into six categories instead of the classical
"light," "moderate," and "heavy" described by Taylor
et al. (14).
An intensity threshold for physical activity in terms
of its association with serum lipids has been observed:
Physical activities with an intensity above 7 kcal/
minute are associated with higher HDL cholesterol
and a lower atherogenic index. This threshold value
coincides with that found by Morris et al. (24, 25) to
be related to the protective effect of physical activity
against coronary heart disease. On the other hand,
physical activities with an intensity above 9 kcal/
minute are associated with lower total cholesterol,
non-HDL cholesterol, and triglyceride levels.
Our participants were seldom involved in physical
activities with an intensity of 9.5-11 kcal/minute (table 1); these physical activities were not significantly
associated with the lipid profile in the bivariate analAm J Epidemiol
Vol. 143, No. 6, 1996
Physical Activity, Fitness, and Lipids
567
TABLE 4. Results of multiple linear regression analysis (p coefficient ± standard error) of the relations of serum lipids and
physical fitness (V02max) with energy expenditure in leisure time physical activity, age, alcohol consumption, and smoking:
Barcelona, Spain, 1991-1993f
Total
cholesterol
(mg/dl)
Non-HDL
cholesterolt
(mg/dl)
HDL
cholesterolt
(mg/dl)
Log
triglycerides
Atherogenic
indext
V0 2 maxt
(ml/kg/minute)
EEPAt (100 kcal/day)
EEPAs^
EEPA.
EEPA . . .
Aqe (years)
Alcohol (g/day)
Current smoking (yes = 1/
no = 0)
Constant
-2.877
-4.535
-0.976
-0.465
-6.729
-2.247
1.329
0.237
±1.837
± 2.739
±2.118
± 2.601
± 3.409*
± 0.780*
±0.192*
± 0.092*
0.024
-0.004
0.058
2.387
2.536
2.052
0.120
0.057
-2.890 ±1.906
-4.460 ± 2.842
-0.983 ± 2.200
-2.813 ±2.699
-9.215 ± 3.538*
-4.279 ± 0.810*
1.217 ±0.200*
0.182 ±0.095
± 0.573
± 0.854
± 0.661
±0.811*
±1.063*
± 0.243*
±0.060*
± 0.029*
4.259 ±2.150*
153.046 ±9.675*
-2.004 ±0.671*
42.736 ± 3.038*
6.330 ± 2.234*
109.670 ± 10.108*
r= 0.433
ft = 0.187
/•= 0.449
ft = 0.202
r = 0.481
ft = 0.231
-0.0138
-0.0129
0.0110
-0.0123
-0.0407
-0.0281
0.0037
0.0017
± 0.0097
±0.0145
±0.0112
±0.0137
±0.0180*
±0.0041*
±0.0010*
±0.0005*
-0.060 ± 0.072
-O.070± 0.107
-0.004 ± 0.083
-0.241 ±0.102*
-0.324 ±0.133*
-0.224 ±0.031*
0.017 ±0.007*
0.000 ± 0.004
0.0476 ±0.0114*
1.7533 ±0.0519*
0.313 ±0.084*
3.834 ±0.381*
r = 0.504
ft = 0.254
! = 0.469
ft = 0.220
-0.866
0.212
0.867
1.024
0.560
1.912
-0.264
±0.331*
±0.478
± 0.369*
±0.454*
± 0.589
±0.137*
± 0.034*
-1.939 ±0.682*
51.730 ±1.538*
r = 0.698
ft = 0.487
* p < 0.05.
t To convert values for total cholesterol, HDL cholesterol, and non-HDL cholesterol to mmol/liter, multiply by 0.02586.
t HDL cholesterol, high density lipoprotein cholesterol; non-HDL cholesterol, total cholesterol minus HDL cholesterol; atherogenic index, total cholesterohHDL
cholesterol; V02max, maximal oxygen uptake (indirect calculation); EEPA, daily average energy expenditure in leisure time physical activity of different intensities:
2-3 kcal/minute, 3.5-5 kcal/minute, 5.5-7 kcal/minute, 7.5-9 kcal/minute, 9.5-11 kcal/minute, and 11.5-12 kcal/minute.
ysis (table 3). Thus, the statistically significant association found between EEPA9 5_x t and the lipid profile
in the multiple linear regression models provides further strength for this relation.
The extent to which our findings may be due to the
acute effect of exercise was assessed by studying the
adjusted effect on the lipid values of physical activity
performed during the week prior to examination. Coefficients obtained by this analysis were not signifi-
cantly different from those of the present models (results not shown). In addition, running was the most
frequent physical activity performed by the marathon
runners. Running was included in the EEPA 1 ] 5 _ ] 2
intensity category. However, physical activities in
intensity categories EEPA 75 _ 9 and EEPA 9 5 _ n were
also associated with lipid levels; their adjusted coefficients were similar to those for the EEPA n 5_12 category.
TABLE 5. Results of multiple linear regression analysis (p coefficient ± standard error) of the relations of serum lipids with
physical fitness (V02max), energy expenditure in leisure time physical activity, age, alcohol consumption, and smoking:
Barcelona, Spain, 1991-1993t
Total
cholesterol
(mg/dl)
V02max$ (ml/kg/minute)
EEPAt (100 kcal/day)
EEP/^3
EEPA^
EEPAs^7
EEPA^
EEP/W,,
EEPA,,^, 2
Age (years)
Alcohol (g/day)
Current smoking
(yes = 1/no = 0)
Constant
-0.357
-2.928
-4.372
-0.694
-0.302
-6.821
-1.616
1.241
0.237
4.011
178.239
± 0.269
±1.939
± 2.780
±2.163
± 2.653
± 3.430*
±0.956
±0.208*
± 0.094*
± 4.053
±16.619*
r = 0.424
n= 0.180
HDL
cholesterolt
(mg/dl)
0.193
-0.200
-0.203
0.054
2.332
2.407
1.688
0.159
0.062
-3.435
30.623
±0.082*
± 0.589
± 0.845
± 0.658
± 0.807*
± 1.043*
±0.291*
±0.063*
± 0.029*
±1.233*
± 5.066*
r= 0.477
n = 0.228
Non-HDL
cholesterolt
(mg/dl)
-0.656 ± 0.403
-4.142 ±2.899
-5.248 ±4.157
-2.730 ± 3.234
-4.138± 3.967
-11.084 ±5.129*
-3.929 ±1.430*
0.695 ± 0.311 *
0.085 ± 0.141
3.446 ± 6.060
176.158 ±24.850*
/•= 0.344
r*= 0.119
Log
triglycerides
-0.0043
-0.0167
-0.0133
0.0118
-0.0080
-0.0392
-0.0187
0.0027
0.0016
0.0691
2.0414
±0.0014*
±0.0100
±0.0143
±0.0112
± 0.0137
± 0.0177*
± 0.0049*
±0.0011*
±0.0005*
± 0.0210*
± 0.0865*
r= 0.507
r*= 0.258
Atherogenic
index}
-0.025
-0.071
-0.064
0.001
-0.228
-0.312
-0.171
0.012
0.000
0.524
5.482
± 0.010*
± 0.076
±0.108
± 0.084
±0.103*
±0.134*
±0.037*
±0.008
± 0.004
±0.158*
± 0.650*
/"= 0.480
P = 0.230
* p < 0.05.
t To convert values for total cholesterol, HDL cholesterol, and non-HDL cholesterol to mmol/liter, multiply by 0.02586.
t HDL cholesterol, high density lipoprotein cholesterol; non-HDL cholesterol, total cholesterol minus HDL cholesterol; atherogenic index,
total cholesterohHDL cholesterol; VO2max, maximal oxygen uptake (indirect calculation); EEPA, daily average energy expenditure in leisure
time physical activity of different intensities: 2-3 kcal/minute, 3.5-5 kcal/minute, 5.5-7 kcal/minute, 7.5-9 kcal/minute, 9.5-11 kcal/minute and
11.5-12 kcal/minute.
Am J Epidemiol
Vol. 143, No. 6, 1996
568
Marrugat et al.
The association found between amount of physical
activity and the lipid profile above the intensity threshold was linear for total cholesterol, HDL cholesterol,
non-HDL cholesterol, and the atherogenic index, and
was logarithmic for triglycerides.
The simultaneous inclusion of EEPA and physical
fitness (VO2max) measurements in the regression
models failed to significantly change the estimated
multiple linear regression coefficients or to improve
the models. It is likely that such a finding is due to the
strong correlation between physical activity and
VO2max (table 3) and to the fact that the physical
activity intensity threshold for lipid modification (>7
kcal/minute) is different from that for V0 2 max (>5
kcal/minute).
The independent protective effect of HDL cholesterol on the risk of coronary heart disease, especially
the total cholesterol: HDL cholesterol ratio, is well
known (26). A direct association between amount of
physical activity and HDL cholesterol levels has been
described (6). An increase of 0.054 mmol/liter (2.09
mg/dl) in HDL cholesterol was associated with every
average 100 kcal/day of energy expended in physical
activities with an intensity greater than 7 kcal/minute
during the previous year. The relation of total EEPA,
regardless of intensity, to HDL cholesterol fell to
0.044 mmol/liter (1.70 mg/dl) for each 100 kcal/day of
average energy expenditure in the past year. Recently,
Lakka and Salonen (20) showed that when total EEPA
was considered, every 100-kcal/day average expenditure in the past week was associated with a 0.017mmol/liter (0.67-mg/dl) increase in HDL cholesterol.
The magnitude of the association was higher in the
present study, probably because total EEPA in our
participants included a greater proportion of highintensity physical activity (EEPA >7 kcal/minute)
(due to our sampling procedure).
It is well known that diet influences serum lipid
levels (27). The inclusion of dietary saturated and
polyunsaturated fatty acid intakes, as well as their
ratio, failed to improve our models. This finding may
be related to the difficulty of achieving an accurate
assessment of diet by questionnaire (28), although a
lack of effect of diet on HDL cholesterol has already
been found by other researchers (29).
Results from models with and without adjustment
for BMI indicate that the association between EEPA
and serum lipids is independent of that of BMI. Consequently, BMI should not be considered the only
mechanism of the beneficial relation between physical
activity and the serum lipid profile.
The physical activity intensity associated with better
physical fitness is lower (5 kcal/minute) than that
associated with the serum lipid profile, and this value
agrees with the recommendation of the American College of Sports Medicine (11). The unexpected negative
relation between EEPA2_3 and V0 2 max may be related to underestimation of physical activities of low
intensity (e.g., walking for pleasure), especially by
very active participants, as has been recently described
by Richardson et al. (30).
An increase of 0.015 mmol/liter (0.6 mg/dl) in HDL
cholesterol for each 10 g/day of alcohol consumption
was observed (table 5). Such a direct relation agrees
with recent findings suggesting that moderate alcohol
consumption has health benefits (31). Smoking had a
negative relation with both physical fitness and lipid
profile (table 4).
In conclusion, recommendations of exercise should
take into account the fact that physical activities such
as walking for pleasure, brisk walking, jogging, and
gardening, all of which have an intensity below 7.5
kcal/minute, are probably insufficient to achieve a
relevant change in the lipid profile. Thus, we believe
that recommending physical activities with intensities
greater than or equal to 7.5 kcal/minute (e.g., climbing
stairs, running faster than 5 miles (8 km) per hour, or
playing tennis or basketball) would be more appropriate for improving not only physical fitness but also the
lipid profile.
ACKNOWLEDGMENTS
This study was supported by grant 90/0668 from the
Fondo de Investigaciones Sanitarias.
The authors thank Professor J. N. Morris for his valuable
comments on the manuscript, David McFarlane for English
revision, and Dr. Marta Pulido for editorial assistance. The
authors appreciate the enthusiastic collaboration of all participants.
The MARATHOM Investigators: A. Arquer, X. Castells,
M. I. Covas, T. De Flores, R. Elosua, X. Fornells, J.
Marrugat, S. Martin, L. Molina, M. Perea, S. Pons, E. Pujol,
P. Pujol, J. A. Richart, J. RubieVPrat, and A. Tomas.
REFERENCES
1. Powell KE, Thompson PD, Caspersen CJ, et al. Physical
activity and the incidence of coronary heart disease. Annu Rev
Public Health 1987;8:253-87.
2. Berlin JA, Colditz GA. A meta-analysis of physical activity in
the prevention of coronary heart disease. Am J Epidemiol
1990;132:612-28.
3. Lakka TA, Venalainen JM, Rauramaa R, et al. Relation of
leisure-time physical activity and cardiorespiratory fitness to
the risk of acute myocardial infarction. N Engl J Med 1994;
330:1549-54.
4. Hambrecht R, Niebauer J, Marburger C, et al. Various intensities of leisure time physical activity in patients with coronary
artery disease: effects on cardiorespiratory fitness and progresAm J Epidemiol
Vol. 143, No. 6, 1996
Physical Activity, Fitness, and Lipids
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
sion of coronary atherosclerotic lesions. J Am Coll Cardiol
1993;22:468-77.
Pekkanen J, Marti B, Nissinen A, et al. Reduction of premature mortality by high physical activity: a 20-year follow-up of
middle-aged Finnish men. Lancet 1987; 1:1473-7.
Haskell WL. Exercise-induced changes in plasma lipids and
lipoproteins. Prev Med 1984,13:23-36.
US Public Health Service, US Department of Health and
Human Services. Year 2000 health objectives for the nation.
Washington, DC: US GPO, 1990.
Roberts WC. An agent with lipid-lowering, antihypertensive,
positive inotropic, negative chronotropic, vasodilating, diuretic, anorexigenic, weight-reducing, cathartic, hypoglycemic, tranquilizing, hypnotic and antidepressive qualities. (Editorial). Am J Cardiol 1984;53:261-2.
Blackburn H, Jacobs DR Jr. Physical activity and the risk of
coronary heart disease. (Editorial). N Engl J Med 1988;319:
1217-19.
Koplan JP, Caspersen CJ, Powell KE. Physical activity, physical fitness, and health: time to act. (Editorial). JAMA 1989;
262:2437.
American College of Sports Medicine position stand. The
recommended quantity and quality of exercise for developing
and maintaining cardiorespiratory and muscular fitness in
healthy adults. Med Sci Sports Exerc 1990;22:265-74.
Blair SN, Kohl HW, Gordon NF, et al. How much physical
activity is good for health? Annu Rev Public Health 1992;13:
99-126.
Paffenbarger RS Jr, Wing AL, Hyde RT. Physical activity as
an index of heart attack risk in college alumni. Am J Epidemiol 1978;108:161-75.
Taylor HL, Jacobs DR Jr, Schucker B, et al. A questionnaire
for the assessment of leisure time physical activities. J Chronic
Dis 1978;31:741-55.
Elosua R, Marrugat J, Molina L, et al. Validation of the
Minnesota Leisure Time Physical Activity Questionnaire
among Spanish men. The MARATHOM Investigators. Am J
Epidemiol 1994;139:1197-209.
Frick MH, Elo O, Haapa K, et al. Helsinki Heart Study:
primary-prevention trial with gemfibrozil in middle-aged men
with dyslipidemia. Safety of treatment, changes in risk factors,
and incidence of coronary heart disease. N Engl J Med 1987;
317:1237-45.
Bruce RA, Kusumi F, Hosmer D. Maximal oxygen intake and
nomographic assessment of functional aerobic impairment in
Am J Epidemiol
Vol. 143, No. 6, 1996
569
cardiovascular disease. Am Heart J 1973;85:545-62.
18. American College of Sports Medicine. Guidelines for graded
exercise testing and exercise prescription. Philadelphia, PA:
Lea and Febiger, 1988:395-9.
19. Foster C, Jackson AS, Pollock ML, et al. Generalized equations for predicting functional capacity from treadmill performance. Am Heart J 1984; 107:1229-34.
20. Lakka TA, Salonen JT. Physical activity and serum lipids: a
cross-sectional population study in eastern Finnish men. Am J
Epidemiol 1992;136:806-18.
21. De Backer G, Komitzer M, Sobolski J, et al. Physical activity
and physical fitness levels of Belgian males aged 40-55 years.
Cardiology 1981;67:110-28.
22. Slattery ML, Jacobs DR Jr, Nichaman MZ. Leisure time
physical activity and coronary heart disease death: The US
Railroad Study. Circulation 1989;79:304-ll.
23. LaPorte R, Brenes G, Dearwater S, et al. HDL cholesterol
across a spectrum of physical activity from quadriplegia to
marathon running. (Letter). Lancet 1983;1:1212-13.
24. Morris JN, Everitt MG, Pollard R, et al. Vigorous exercise in
leisure-time: protection against coronary heart disease. Lancet
1980;2:1207-10.
25. Morris JN, Clayton DG, Everitt MG, et al. Exercise in leisure
time: coronary attack and death rates. Br Heart J 1990;63:
325-34.
26. Gordon DJ, Probstfield JL, Garrison RJ, et al. High-density
lipoprotein cholesterol and cardiovascular disease: four prospective American studies. Circulation 1989;79:8-15.
27. Keys A, Anderson JT, Grande F. Prediction of serum-cholesterol responses of man to changes in fats in the diet. Lancet
1957;2:959-66.
28. Sabat6 J. Estimaci6n de la ingesta diete'tica: metodos y desafios. Med Clin (Bare) 1993;100:591-6.
29. Hartung GH, Foreyt JP, Mitchell RE, et al. Relation of diet to
high-density-lipoprotein cholesterol in middle-aged marathon
runners, joggers, and inactive men. N Engl J Med 1980;302:
357-61.
30. Richardson MT, Leon AS, Jacobs DR Jr, et al. Comprehensive
evaluation of the Minnesota Leisure Time Physical Activity
Questionnaire. J Clin Epidemiol 1994;47:271-81.
31. Suh I, Shaten J, Cutler JA, et al. Alcohol use and mortality
from coronary heart disease: the role of high-density lipoprotein cholesterol. The Multiple Risk Factor Intervention Trial
Research Group. Ann Intern Med 1992;116:881-7.