Women
How Much Is
Walking for Health and
Fitness
Enough?
John J. Duncan, PhD; Neil F. Gordon, MBBCh, PhD; Chris B. Scott, MS
Objective.\p=m-\Westudied whether the quantity and quality of walking necessary to decrease the risk of cardiovascular disease among women differed
substantially from that required to improve cardiorespiratory fitness.
Design.\p=m-\Arandomized, controlled, dose-response clinical trial with a
follow-up of 24 weeks.
Setting.\p=m-\Aprivate, nonprofit biomedical research facility.
Participants.\p=m-\One hundred two sedentary premenopausal women, 20 to 40
years of age, were randomized to one of four treatment groups; 59 completed
the study (16 aerobic walkers [8.0-km/h group], 12 brisk walkers [6.4-km/h
group], 18 strollers [4.8-km/h group], and 13 sedentary controls). Eighty-one
percent were white, 17% black, and 2% Hispanic.
Intervention.\p=m-\Interventiongroups walked 4.8 km per day, 5 days per week at
8.0 km/h, 6.4 km/h, or 4.8 km/h on a tartan-surfaced, 1.6-km track for 24 weeks.
Main Outcome Measures.\p=m-\Fitness(determined by maximal oxygen uptake) and cardiovascular risk factors (determined by resting blood pressure and
serum lipid and lipoprotein levels).
Results.\p=m-\Ascompared with controls, maximal oxygen uptake increased
significantly (P<.0001) and in a dose-response manner (aerobic walkers>brisk
walkers>strollers). In contrast, high-density lipoprotein cholesterol concentrations were not dose related and increased significantly (P<.05) and to the same
extent among women who experienced considerable improvements in their
physical fitness (8.0-km/h group, +0.08 mmol/L) and those who had only minimal improvements in fitness (4.8-km/h group, +0.08 mmol/L). High-density lipoprotein cholesterol also increased among the 6.4-km/h group, but did not attain statistical significance (+0.06 mmol/L; P=.06). Dietary patterns revealed
no significant differences among groups.
Conclusion.\p=m-\Thus,we conclude that vigorous exercise is not necessary for
women to obtain meaningful improvements in their lipoprotein profile. Walking
at intensities that do not have a major impact on cardiorespiratory fitness may
nonetheless produce equally favorable changes in the cardiovascular risk profile.
(JAMA. 1991;266:3295-3299)
FREQUENCY, intensity, and duration
provide the framework for
developing an exercise prescription. The
of exercise
interaction of these factors has been
thoroughly examined only with regard
to the proper dose of exercise required
to increase cardiorespiratory fitness.1
Because cardiorespiratory fitness and
cardiovascular health have been consid¬
ered synonymous, it is not surprising
that the guidelines published in 1978 by
the American College of Sports Medi¬
cine describing the quality and quantity
of exercise necessary to increase fitness
are often extrapolated to the prescrip¬
tion of exercise to prevent cardiovas¬
cular disease.2 However, attainment of
From the Division of Exercise Physiology, The Cooper Institute for Aerobics Research, Dallas, Tex.
Reprint requests to The Cooper Institute for Aerobics
Research, 12330 Preston Rd, Dallas, TX 75230 (Dr
Duncan).
a
fit state may not be necessary to mod¬
ify specific cardiovascular disease risk
factors favorably.
The latter hypothesis is supported by
recent epidemiologie data that demon¬
strate that persons who participate in
physical activities that are less intense
and, therefore, unlikely to have a pro¬
found impact on cardiorespiratory fit¬
ness, do, nonetheless, derive cardiovas¬
cular health benefits.3·4 However, it is
difficult to quantify the amount of phys¬
ical activity required to provide an ap¬
parent protective effect against the de¬
velopment of cardiovascular disease
from epidemiologie data alone. While
ample support shows low-level activity
may lead to a less atherogenic lipid pro¬
file among men,5 few studies have in¬
vestigated this possibility among
women.
Therefore,
independently the relationship between
cardiorespiratory fitness and cardiovas¬
cular health among women, we designed
a 24-week, dose-response, randomized
clinical trial among sedentary, premenopausal women, in which walking exer¬
cise intensity varied across three treat¬
(strollers, brisk walkers,
and aerobic walkers) while keeping the
distance and frequency constant. Sta¬
tistical analyses focused on whether
changes in clinical cardiovascular risk
factors paralleled changes in cardiores¬
piratory fitness across the three walk¬
ing groups and the control group.
ment groups
SUBJECTS AND METHODS
More than 300 women who responded
to various media sources calling for vol¬
unteers to participate in this study were
interviewed by telephone. One hundred
two
adult, premenopausal, sedentary
women
were
who
were
20 to 40 years of age
randomly selected from this pool
of volunteers if they (1) did not smoke;
(2) consumed fewer than three alcoholic
drinks per day; (3) did not currently
receive dietary interventions; (4) did not
have cardiopulmonary and/or musculoskeletal diseases; and (5) did not reg¬
ularly exercise more than 1 day per week
for the previous 6 months. Women with
resting blood pressures below 160/90 mm
Hg, total serum cholesterol levels be¬
low 6.59 mmol/L, and serum triglycér¬
ide levels below 2.25 mmol/L who were
willing to accept random assignment to
any of the four treatment groups were
entered into the study, randomized to
intervention groups, and followed up for
24 weeks. Written informed consent was
obtained from each participant prior to
entry into the study. All groups were
advised not to change dietary, exercise
(other than prescribed), or other life¬
style habits. Three-day dietary food
records and physical activity question¬
naires were administered at baseline and
again at the end of the 24-week inter¬
vention study to evaluate whether par¬
ticipants adhered to these recommen¬
dations.
Control subjects (n 21) remained
sedentary for the duration of the study
and were not contacted except to sched¬
ule follow-up testing. Subjects assigned
to the aerobic walkers (n 29), brisk
walkers (n 26), or strollers (n 26)
=
=
to
explore separately
and
=
Downloaded from www.jama.com at University of Colorado Denver HSL on August 17, 2009
=
trained under the supervision of an ex¬
ercise physiologist on a tartan-surfaced,
1.6-km track, 5 days per week, for 24
weeks.
The frequency, intensity, and dura¬
tion of each exercise session was stan¬
dardized across all three walking inter¬
vention groups. Initially, each group
walked a distance of 2.4 km, 5 days per
week. Thereafter, walking distance was
gradually increased each week until all
subjects reached a maintenance distance
of 4.8 km by the seventh week of train¬
ing. Similarly, walking intensity
was
standardized to the same relative level
each week among each of three walking
groups. Initial exercise intensities were
equivalent to 70% of group assignment
(ie, 70% of 8.0 km/h [aerobic walkers],
70% of 6.4 km/h [brisk walkers], and
70% of 4.8 km/h [strollers]) and subse¬
quently increased to 100% of prescribed
intensity by the 14th week of training
and maintained until completion of the
study (week 24).
Resting blood pressure, serum lipid
and lipoprotein levels, maximal oxygen
uptake (Vo2 max), nude body weight,
and percentage of body fat were mea¬
sured at baseline and after 24 weeks of
intervention. Resting blood pressure
was measured in triplicate in the seated
position on three separate days accord¬
ing to recommendations of the Ameri¬
can Heart Association.6 The same
trained observer made all measurements
with a mercury sphygmomanometer.
The mean of the three readings on the
third day was used as baseline and fol¬
low-up values. Systolic and diastolic
blood pressures were recorded at the
first and fifth phases of Korotkoffs
sounds.
Blood specimens were collected in the
morning after abstaining from all food,
beverages (except water), and vigorous
activity for 12 to 14 hours. Serum was
separated from venous blood within 30
minutes by centrifugation at 3400 rpm
for 10 minutes and stored at -80°C until
analyses. Total serum cholesterol and
triglycéride levels were measured us¬
ing an Olympus AU 5000 (Olympus Corp,
Lake Success, NY) according to enzy¬
matic procedures. High-density lipopro¬
tein (HDL) cholesterol levels were mea¬
sured after very-low-density lipoprotein
and low-density lipoprotein (LDL) cho¬
lesterol levels were precipitated using a
modified phosphotungstic acid (0.55
mmol/L) and magnesium chloride (25
mmol/L) reagent.7 Four hundred microliters of phosphotungstic acid and mag¬
nesium chloride solution (6.4 g of phos¬
photungstic acid, 20 mL of magnesium
chloride in 3 L of deionized water) were
added to 200 µL of serum. After the
solution was mixed, the suspension was
centrifuged at 3200 rpm at 4°C and an¬
alyzed. This method, which has been
demonstrated to be more stable than
concentrated phosphotungstate re¬
agents, markedly reduces the interfer¬
ence from high-serum triglycéride con¬
centrations. The LDL cholesterol was
calculated with the following equation:
total cholesterol (HDL cholesterol +
[triglycerides/5]).8
-
Blinded control pools from the Cen¬
ters for Disease Control, Atlanta, Ga,
yielded values within 1.9% to 3.5% of
Centers for Disease Control target val¬
ues. The average day-to-day standard
deviation (SD) and coefficient of varia¬
tion (CV) of pooled control serum sam¬
ples over the duration of the study were
2.5 SD, 1.5% CV for cholesterol; 1.1 SD,
2.8% CV for HDL cholesterol; and 5.0
SD, 3.5% CV for triglycérides. Calibra¬
tor samples (BMD Preciset,
Boehringer
Diagnostics, Indianapolis,
Ind) were run with each patient sample
and checked for linearity. There was no
Mannheim
evidence of drift within or between any
of the analyses.
Baseline and posttraining serum lipid
and lipoprotein levels were calculated
using the mean of samples drawn on two
or three different days (3 days if vari¬
ation in duplicate samples exceeded 25%
for HDL cholesterol). Baseline and posttraining lipid and lipoprotein samples
were taken at the same phase of the
menstrual cycle.
Body weight and height were mea¬
sured anthropometrically with an Acme
seated scale (model ASCMIN, Acme
Scale Company, Oakland, Calif) and an
anthropometer accurate to the nearest
0.01 kg and 0.1 cm, respectively. Body
density was determined by hydrostatic
weighing techniques, and percentage of
body fat was calculated according to the
Siri equation.9
Maximal graded treadmill testing us¬
ing automated cardiorespiratory moni¬
toring techniques (MMC Horizon Sys¬
tem Exercise Evaluation Cart, Sensormedics, Yorba Linda, Calif) was per¬
formed on all subjects using a modified
Balke protocol.10 Gas and volume cali¬
brations were performed before and af¬
ter each test. Subjects exercised until
volitional exhaustion, and the peak ox¬
ygen uptake and heart rate attained
were taken as Vo2 max and maximal
heart rate, respectively.
Statistical analyses were performed
using the Statistical Analysis System.11
Power and sample-size determinations
were calculated based on an effect size
of 7 mL/kg per minute (Vo2 max), and
an SD of 3 mL/kg per minute. An esti¬
mated sample size of 25 subjects per
group at an level of .05 yielded a 1-ß
level of .99. Dropouts during the study
lowered the power to an approximate
level of .05 at a 1-ß level of .94. Group
randomization was performed by the
principal investigator via a table of ran¬
dom numbers. Initially, 21 subjects were
randomly assigned to each of the four
groups. We continued random assign¬
ment to walking groups only until each
of the three walking groups comprised
26 subjects. Three subjects were added
to the aerobic walkers in anticipation of
higher intensity-related dropout rates
among the faster walkers. All baseline
and posttest-dependent variables (ex¬
cept percentage of body fat) were mea¬
sured by personnel who were blinded to
individual participant group assign¬
ments.
Exact probability and analysis of vari¬
ance were used to compare
demographic
data (age and race) and baseline depen¬
dent variables (Vo2 max, body compo¬
sition, lipid levels, and resting blood
pressure) for the four treatment groups.
Changes in group scores in cardiorespiratory fitness, body composition, and
cardiovascular health-related variables
were analyzed for subjects who com¬
pleted the study (aerobic walkers
[n 16], brisk walkers [n 12], strollers
[n 18], and controls [n 13]) by an anal¬
ysis of variance model. Changes within
the treatment groups were performed
=
=
=
=
using paired t tests. Group contrasts
performed with a Scheffe test to
guard against experimental error rate
of unequal-sized samples. Statistical sig¬
were
nificance was established at P<. 05. Data
are presented as the mean+SD.
RESULTS
Demographic and Compliance Data
Table 1 shows that cardiorespiratory
and cardiovascular health-related vari¬
ables were comparable at baseline across
all four treatment groups. Eighty-one
percent of the women who participated
in the study were white, 17% black, and
2% Hispanic. Five women in the aerobicwalkers group, seven women in the
brisk-walkers group, five women in the
strollers group, and four women in the
control group were taking birth control
pills. Compliance to training (total num¬
ber of sessions attended divided by total
number of sessions possible) exceeded
85% for all three walking groups.
Dropout rates were as follows: four
in the aerobic-walkers group,
in the strollers group, and one in the
control group became pregnant during
the study and subsequently withdrew
from further participation. In addition,
nine aerobic walkers, 14 brisk walkers,
seven strollers, and seven controls
dropped out due to relocation, medical
reasons unrelated to the study, or loss
of interest. Finally, three aerobic walkwomen
one
Downloaded from www.jama.com at University of Colorado Denver HSL on August 17, 2009
(two receiving thyroid medication
one with endometriosis) and three
women in the control group (two receiv¬
ers
Table 1.—Baseline Values
(Mean ± SD) for Study Groups*
and
ing thyroid medication and one receiv¬
ing steroid hormone therapy after base¬
line testing) were excluded from lipid
level and/or blood pressure analyses.
Statistical analyses compared drop¬
outs with subjects who completed the
study. No differences were observed be¬
tween dropouts and corresponding group
members who completed the study re¬
garding fitness (Vo2 max), body fat, or
lipid level concentrations (P>.05). Drop¬
outs in the groups of brisk walkers,
strollers, and controls did have signifi¬
cantly lower (P<.001) blood pressure.
Dropouts in the brisk-walkers group also
had significantly higher (P<.001) body
weight compared with subjects who com¬
pleted the study.
Walking for Fitness
Table 2 shows Vo2 max and body com¬
position responses to 24 weeks of walk¬
ing. A linear, dose-response gradient
across all walking groups was observed
for Vo2 max, with the aerobic walkers
experiencing the greatest increase of
Vo2 max (+5.0 mL/kg per minute), the
brisk walkers experiencing a moderate
rise of Vo2 max (+3.0 mL/kg per
minute), and the strollers achieving a
minimal increase in Vo2 max (+1.4 raL/
kg per minute). Posttraining Vo2 max
was not significantly changed among
subjects in the control group. Analysis
of group differences of Vo2 max yielded
a significant overall F test (F
16.2 [3,51
dfl; P<.001). Group contrasts (Scheffe)
revealed that Vo2 max differed signifi¬
cantly (P<.05) between the control
group and each of the walking groups
=
and between the strollers and aerobic
walkers.
Heart rates recorded during exercise
training mirrored the dose-response
Vo2 max gradient observed across all
three walking groups. The aerobic walk¬
ers exercised at 86% (163 beats per
minute) of their maximal heart rate,
while the brisk walkers and strollers
trained at 67% (126 beats per minute)
and 56% (106 beats per minute) of their
maximal heart rates.
Walking for Health
Table 3 shows changes in serum lipid
and lipoprotein levels and blood pres¬
sure after 24 weeks of training. Changes
in total cholesterol, LDL cholesterol,
HDL cholesterol, and serum triglycér¬
ides did not differ significantly between
any of the treatment and control groups.
However, mean serum concentrations
of HDL cholesterol (Figure) increased
significantly (P<.05) among the aerobic
walkers (+0.08 mmol/L) and strollers
Brisk
Walkers
(n 16)
Strollers
(n 12)
(n 18)
163.2 ±6.4
165.0 ±8.3
60.3 ±9.5
27.5 ±7.6
64.2 ±3.8
26.2 ±6.6
164.9 ±4.8
62.0 ±9.8
841.0 ±153
902.0 ± 222
Aerobic
Walkers
Variable
=
Height, cm
Weight, kg
Body fat, %
Maximum oxygen
Treadmill time, s
consumption, ml_/kg
Control
=
27.9 ±6.0
32.1+6.8
870.0 ± 239
per min
Total cholesterol, mmol/Lf
LDL cholesterol, mmol/Lf
4.68 ± 0.74
4.93 ± 0.87
2.87 ±0.67
2.99 ±0.69
HDL cholesterol, mmol/Lt
1.42 ±0.43
0.89 ±0.36
1.53 ±0.36
0.92 ±0.32
105/70 ±8/7
109/74 ±9/8
Triglycérides, mmol/LtResting seated blood pressure,
*No statistical differences
mm
were
Hg
1.10±0.51
108/73 ±6/9
observed between any of the four treatment groups. LDL indicates
(Mean±SD)
in
Baseline
Trained
Change, %
consumption, mL/kg per min
13)_29.3 ±4.9_27.6 5.1_-6
17)_31.8±6.9_33.2±6.6_+4f
12)_32.4±6.1_35.4 ±60_+9t
Control group (
Strollers (n
Brisk walkers (n
=
Aerobic walkers
Treadmill time,
low-density
Fitness, Body Weight, and Body Composition After 24 Weeks of
Variable
Maximum oxygen
29.3 ±4.9
2.98 ±0.79
1.34 ±0.36
lipoprotein; HDL, high-density lipoprotein.
tTo convert to milligrams per deciliter, multiply by 38.67.
tTo convert to milligrams per deciliter, multiply by 88.57.
Table 2—Changes
Intervention*
Group
(n 13)
=
=
+
=
=
(n 13)
30.6 ±3.9
=
35.6 + 4.7
+16tt
s
Control group ( 13)_776 ±181_692 ±172_-11
Strollers ( 18)_870 ± 239_936 ±221_+8t
Brisk walkers (n 12)_902 ± 222_1021 ±199_+13t
Aerobic walkers (n 16)
841 ±153
997 ±167
+19tt
=
=
=
=
Body weight, kg
Control group
66.5 ±14.3
+6
(n 13)
70.7 + 18.0
62.0 ±9.8
+1
Strollers (n 18)
62.8 + 10.2
Brisk walkers (n 12)_64.2 ±3.8_64.3
±4.0_0§
Aerobic walkers (n 16)
60.3 ±9.5
+2
61.4 ±10.4
=
=
=
=
Body fat, %
Control group (n= 13)_32.2±6.1_33.5 + 8.1_+4
Strollers (n 18)_27.9 ±6.0_26.2 ±6.8_-6§
=
Brisk walkers (n
Aerobic walkers
=
12)_26.2±6.6_24.8±6.2_-5
27.5 ±7.6
(n 16)
=
26.4 ±7.5
-4
»Strollers travel 4.8 km/h; brisk walkers, 6.4 km/h; and aerobic walkers, 8.0 km/h.
tChange from baseline is different from controls (P< .001 ).
íChange from baseline is different from strollers {P<.001 ).
§Change from baseline is different from controls (P<.05).
(+0.08 mmol/L) but
not in the brisk
walkers (+ 0.06 mmol/L; =.06) or con¬
trols ( + 0.02 mmol/L;
.77). The ratio
of total cholesterol and HDL cholesterol
was lowered significantly (P<.01) within
the strollers, but not in any of the other
three treatment groups. Self-reported
alcohol intake, caloric intake (assessed
by 3-day dietary food records), dietary
fat, carbohydrate, and protein intake,
medication use, and smoking habits did
not change significantly among any of
the four study groups after 24 weeks.
The Figure illustrates the fitness and
lipid responses to 24 weeks of training.
Cardiorespiratory fitness increased in a
dose-response manner with increasing
levels of walking intensity while HDL
cholesterol concentrations increased irre¬
spective of the dose or walking intensity.
Resting seated blood pressure did not
change significantly within or between
=
any of the walking groups after 24 weeks
(Table 3). Similarly,
no
changes
were
observed among subjects in the control
group.
COMMENT
We have attempted to
quantify,
in
a
linear, dose-response manner, the health
and fitness implications of participating
in a popular and frequently prescribed
mode of exercise—walking.12 The
Vo2 max data indicate that for those
who are interested in obtaining mean¬
ingful improvements in cardiorespira¬
tory fitness, walking at a brisk to aer¬
obic pace provides the physiologic stim¬
ulus necessary to accomplish this goal.
The recognized standard of cardiores¬
piratory fitness, Vo2 max, increased in
a linear, dose-response manner from the
strollers (+1.4 mL/kg per minute) to
the brisk walkers (+3.0 mL/kg per
Downloaded from www.jama.com at University of Colorado Denver HSL on August 17, 2009
Table 3.—Changes (Mean ± SD) in
24 Weeks of Intervention*
Resting
Variable
Seated blood pressure, mm
Control group (n 10)
Strollers (n 18)
Brisk walkers (n 12)
Aerobic walkers (n 13)
Total cholesterol, mmol/L
Control group (n 10)
Strollers (n 18)
Brisk walkers (n 12)
Hg
=
=
Blood
Pressure, Lipids, and Lipoprotein Concentrations After
Baseline
Trained
Change, %
108/74 ±7/6
110/75±7/7
105/73±11/6
110/73 ±10/6
105/70 ±7/8
±2/1
108/73 ±6/9
=
=
=
=
=
Aerobic walkers (n 13)
LDL cholesterol, mmol/L
Control group (n= 10)
Strollers (n 18)
Brisk walkers (n 12)
Aerobic walkers (n 13)
HDL cholesterol, mmol/L
Control group (n 10)
Strollers (n 18)
Brisk walkers (n 12)
Aerobic walkers (n 13)
=
=
=
=
=
=
=
=
Triglycérides, mmol/L
Control group (n 10)
Strollers (n 18)
Brisk walkers (n 12)
Aerobic walkers (n 13)
=
=
=
105/70 + 6/7
4.44 ±0.83
4.55 ±0.70
4.73 ±0.94
4.75 ±0.81
4.82 ±0.95
4.93 ±0.87
4.62 ±0.76
=
Strollers (n 18)
Brisk walkers (n 12)
Aerobic walkers (n 13)
=
=
=
0/0
+2
4.85 ±0.79
2.63 ±0.57
2.76 ±0.54
2.98 + 0.79
2.99 ±0.69
2.81 ±0.59
2.86 ±0.77
2.90 + 0.63
+3
1.43 ±0.33
1.34 ±0.36
1.45 + 0.30
+
1.53 ±0.36
1.39 ±0.31
1.59 ±0.34
1.47 + 0.31
+
0.85 0.45
1.10±0.51
0.70 ±0.39
1.01 ±0.57
0.92 ±0.32
0.92 ±0.40
1 05 ±0.52
3.21 ±0.05
3.79 ±1.00
3.37 ±0.78
3.26 ±0.56
3.54 ±0.87
-8t
1
+ 6t
=
Cholesterol and HDL ratio
Control group (n 10)
-3/0
+
6t
-18t
+2
-n
3.13±0.91
3.45 ±0.81
3.52 ±0.96
•Strollers travel 4.8 km/h; brisk walkers, 6.4 km/h; and aerobic walkers, 8 km/h. LDL indicates low-density
lipoprotein; HDL, high-density lipoprotein.
tDifference from baseline (P<05).
Effect of walking intensity on changes from baseline in maximal oxygen uptake (AVo2 max) and
lipoprotein cholesterol ( HDL cholesterol) after 24 weeks of exercise training.
minute) and to the aerobic walkers (+5.0
mL/kg per minute). Furthermore, un¬
other modes of exercise,13
higher intensities did not
come at the expense of a greater risk of
orthopedic injury, as there were no walk¬
ing-related injuries that necessitated
consultation with a physician. Thus,
these results suggest that moderate to
fast walking provides a safe and effec¬
tive means for young to middle-aged
like
some
walking
at
high-density
to achieve significant gains in
cardiorespiratory fitness levels.
A second important finding of our
study involved the relationship between
exercise intensity and HDL cholesterol
concentrations. Our study provides clin¬
ical evidence that the quantity and qual¬
ity of exercise needed to decrease the
risk for developing cardiovascular dis¬
ease may differ substantially from what
is required for improvements in cardio-
women
respiratory fitness. In particular, women
who walked 4.8 km, 5 days per week, at
low intensity for a long duration (4.8
km/h) achieved only a minimal increase
in fitness, but nonetheless experienced
the same significant increase in HDL
cholesterol concentration (+0.08 mmol/
L; P<.05) as women who followed a
similar program but walked vigorously
enough to achieve an aerobically trained
state (8.0 km/h; +0.08 mmol/L; P<.05).
The direction and magnitude of change
in HDL cholesterol among the brisk
walkers (+0.06 mmol/L; =.06) were
similar to the other two walking groups,
but did not reach statistical significance,
possibly due to the smaller sample size.
Clearly, the trend was toward an in¬
crease in HDL cholesterol among all
three walking groups. Thus, our study
is the first to show that within a group
of healthy women the rise in HDL cho¬
lesterol, unlike the rise in fitness, is not
related to intensity of exercise.
Our findings agree with cross-sec¬
tional, epidemiologie data reported by
LaPorte et al14 and more recently by
Härtung et al15 showing HDL choles¬
terol increases across a spectrum of ex¬
ercise intensity—from exercise for quadriplegia to marathon training. Their data
suggest that both low- and high-inten¬
sity activity increases HDL cholesterol
concentrations by the same relative lev¬
els. However, in contrast to epidemio¬
logie data, inconsistent results have been
reported from the few prospective clin¬
ical studies that have investigated the
lipid response to training among women.
Rotkis et al16 observed a 0.13 mmol/L
increase in the HDL cholesterol levels
of 22 women who increased their train¬
ing distance from 26.8 to 74.2 km per
week. Incontrasi, other studies17"20 failed
to demonstrate a significant rise in HDL
cholesterol concentrations after train¬
ing. Inconsistent results in a rise in HDL
cholesterol through exercise have also
been reported.21"23 Goldberg and Elliot21
and Wood et al23 point out that most
studies investigating the lipid response
to training among women have been
short term and may not have allowed
enough time for significant alterations
of lipids and lipoprotéine to occur. Fur¬
thermore, Wood et al23 have reported
that HDL cholesterol is more likely to
increase significantly if the length of
training exceeds 12 weeks (six of eight
longitudinal exercise studies reviewed
that were 13 weeks or longer were as¬
sociated with a significant increase in
HDL cholesterol concentrations) and to
remain unchanged if the length of train¬
ing is 12 weeks or less (HDL cholesterol
was unchanged in 12 of 14 longitudinal
exercise studies that were 12 weeks or
less). These observations may at least
Downloaded from www.jama.com at University of Colorado Denver HSL on August 17, 2009
partly explain the apparent discrepancy
between our study and previous train¬
ing studies of shorter durations.
Upon initial review of our data, the
magnitude of change in HDL choles¬
terol appears modest. However, from a
public health perspective, even small
improvements in coronary risk factors,
if established on a population basis, could
lower cardiovascular-related mortality.
Low to moderate physical activities have
greater compliance rates than more vig¬
orous exercise activities, are more eas¬
ily incorporated into one's daily life-style,
and are well maintained over time. Clin¬
ical data indicate every 1% rise in HDL
cholesterol lowers the risk of coronary
disease by as much as 3%.24 Thus, an
important public health impact may be
obtained simply by persuading the ma¬
jority of the population who are least
active to become just a little more ac¬
tive. In fact, this hypothesis is supported
by a recent epidemiologie report that
suggests that women who regularly par¬
ticipate in physical activities, even at
low levels, may experience lower allcause mortality rates compared with a
cohort of sedentary women.25
The randomized design of our study,
collection of multiple blood samples, and
control for the potential influence of the
menstrual cycle by collecting blood sam¬
ples at the same phase in the menstrual
cycle at baseline and at the 24-week follow-
period, strengthen our findings. Sim¬
ilarly, the fact that changes in HDL choup
lesterol concentrations occurred without
significant concomitant changes in dietary
intake, medication use, or smoking habit
provides additional evidence that the ob¬
served HDL cholesterol responses were
associated with walking rather than other
factors that are known to influence lipopro¬
tein concentrations. Although the mech¬
anisms whereby walking confers its HDL
cholesterol effects are unclear, previous
research has shown that fat loss, whether
induced by dieting or exercise, produces
favorable changes in plasma lipoprotein
levels.26 We found the rise in HDL cho¬
lesterol concentration to be associated
with significant fat loss among the stroll¬
ers, but not among the aerobic walkers.
It remains to be determined whether
the changes in HDL cholesterol concen¬
trations that were observed in the
present study are applicable to popula¬
tions that
are
of a different age, sex,
or
ethnicity than those in the present study.
Similarly, it is unknown whether indi¬
viduals who are at a high risk of devel¬
oping cardiovascular disease or those
who have a cardiovascular disease will
respond in a similar manner as observed
among the healthy women who partic¬
ipated in our study. Finally, it is uncer¬
tain whether health or fitness outcomes
noted in our study would be observed if
other aspects of the exercise prescrip¬
tion (ie, frequency and/or duration of
exercise) were manipulated while hold¬
ing intensity constant. Further research
is needed to clarify these issues.
In summary, our findings indicate that
different walking intensities influence
various aspects of cardiovascular health
and fitness. This observation may serve
as a basis to divide what was previously
one exercise prescription into at least
two—one that follows the American Col¬
lege of Sports Medicine's guidelines with
emphasis on fitness and one designed to
enhance the lipid profile without neces¬
sarily having a major impact on fitness.
In clinical practice, the exercise pre¬
scription will depend on the needs, goals,
abilities, and preferences of each par¬
ticipant. With these objectives in mind, a
sliding scale as to the intensity and amount
of time necessary to increase cardiorespiratory fitness, improve cardiovascular
lipid profile, or both will address the needs
and concerns of each participant, and may
also mean that fewer individuals will fail
to adhere to their prescribed exercise reg¬
imen. From a public health perspective,
it may be more advisable to focus on en¬
couraging the masses to participate reg¬
ularly in low-level activities rather than
advising a few to do more vigorous ex¬
ercise. Perhaps the Greek physician, Hip¬
pocrates (460-377 BC), exhibited great
insight when he wrote "walking is man's
best medicine."
This study was supported by a grant from the
Naturalizer division of Brown Shoe Co, St Louis, Mo.
We thank Karen Rudling, MD, Richard R. Con¬
stant, MD, Carla Sottovia, MS, and Kia Vaandrager for providing extensive and substantive
technical assistance and advice throughout the
course of this study.
References
1. 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;22:265-274.
2. Pollock ML, Froelicher VF. Position stand of
the American College of Sports Medicine: the recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and
muscular fitness in healthy adults. J Cardiopulmonary Rehabil. 1990;10:235-245.
3. Leon AS, Connett J, Jacobs DR Jr, Rauramaa
R. Leisure-time physical activity levels and risk of
coronary heart disease and death: the Multiple Risk
Factor Intervention Trial. JAMA. 1987;258:2388\x=req-\
2395.
4. Paffenbarger RSJr, Hyde RT, Wing AL, Hsieh
CC. Physical activity, all-cause mortality and longevity of college alumni. N Engl J Med. 1986;314:
605-613.
5. Wood PD, Stefanick ML. Exercise, fitness, and
atherosclerosis. In: Bouchard C, Shephard RJ,
Stephens T, Sutton JR, McPherson BD, eds. Exercise Fitness, and Health: A Consensus of Current Knowledge. Champaign, Ill: Human Kinetics;
1990:409-424.
6. American Heart Association. Recommendations
for human blood pressure determination by sphygmomanometers. Dallas, Tex: American Heart Association; 1987.
7. Assman G, Schriewer H, Schmitz G, Hagele E.
Quantification of high-density lipoprotein cholesterol by precipitation with phosphotungstic acid/
MgCl2. Clin Chem. 1983;29:2026-2030.
8. Friedewald WT, Levy RI, Fredrickson DS. Es-
timation of the concentration of low-density lipoprotein cholesterol in plasma without use of preparative ultracentrifuge. Clin Chem. 1972;18:499-502.
9. Siri WE. Body composition from fluid spaces
and density: analysis of methods. In: Brozek J,
Henschel A, eds. Techniques for Measuring Body
Composition. Washington, DC: National Academy
of Sciences; 1961:223-244.
10. Balke B, Ware RW. An experimental study of
physical fitness on air force personnel. US Armed
Forces Med J. 1959;10:675-688.
11. SAS. Version 5. Cary, NC: SAS Institute Inc;
1985.
12. Ryan AJ. Exercise is medicine. Phys Sports
Med. 1983;11:10.
13. Pollock ML, Wilmore JH. Exercise in Health
and Disease: Evaluation and Prescriptionfor Prevention and Rehabilitation. 2nd ed. Philadelphia,
PA: WB Saunders Co; 1990:371-484.
14. LaPorte RE, Brenes G, Dearwater S, et al.
HDL cholesterol across a spectrum of physical activity from quadriplegia to marathon running. Lancet. 1983;1:1212-1213.
15. Hartung GH, Lally DA, Little JL, Goebert
DA. Blood lipids and aerobic fitness in active and
inactive men with disabilities. Med Sci Sports Exerc. 1990;22:S47. Abstract.
16. Rotkis T, Boyden TW, Pamenter RW, Stanforth P, Wilmore J. High-density lipoprotein cholesterol and body composition of female runners.
Metabolism. 1981;30:994-995.
17. Moll ME, Williams RS, Lester RM, Quarfordt
SH, Wallace AG. Cholesterol metabolism in
nonobese women. Atherosclerosis. 1979;34:159-166.
18. Lewis S, Haskell WL, Wood PD, Manoogian
N, Baily JE, Pereira MB. Effects of physical activity on weight reduction in obese, middle-aged
women. Am J Clin Nutr. 1976;29:151-156.
19. Frey MA, Doerr BM, Laubach LL, Mann BL,
Glueck CJ. Exercise does not change high-density
lipoprotein cholesterol in women after 10 weeks of
training. Metabolism. 1982;31:1142-1146.
20. Brownell KP, Bachorik PS, Ayerle RS. Changes
in plasma lipid and lipoprotein levels in men and
after a program of moderate exercise. Circulation. 1982;65:477-484.
21. Goldberg L, Elliot DL. The effect of physical
activity on lipid and lipoprotein levels. Med Clin
North Am. 1985;69:41-55.
22. Haskell WL. The influence of exercise training
on plasma lipids and lipoproteins in health and disease. Acta Med Scand. 1986;711(suppl 711):25-37.
23. Wood PD, Williams PT, Haskell WL. Physical
activity and high-density lipoproteins. In: Miller
NE, Miller GJ, eds. Clinical and Metabolic Aspects of High Density Lipoproteins. New York,
NY: Elsevier Science Publishers; 1984:133-165.
24. Manninen V, Elo MO, Frick MH, et al. Lipid
alterations and decline in the incidence of coronary
heart disease in the Helsinki Heart Study. JAMA.
women
1988;260:641-651.
25. Blair SN, Kohl HW III, Paffenbarger RS Jr,
Clark DG, Cooper KH, Gibbons LW. Physical fitness
and all-cause mortality: a prospective study ofhealthy
men and women. JAMA. 1989;262:2395-2401.
26. Wood PD, Stefanick ML, Dreon DM, et al.
Changes in plasma lipids and lipoproteins in overweight men during weight loss through dieting as
compared with exercise. N Engl J Med. 1988;
319:1173-1179.
Downloaded from www.jama.com at University of Colorado Denver HSL on August 17, 2009