ORIGINAL CONTRIBUTION
Cardiorespiratory Fitness in Young Adulthood
and the Development of Cardiovascular
Disease Risk Factors
Mercedes R. Carnethon, PhD
Samuel S. Gidding, MD
Rodrigo Nehgme, MD
Stephen Sidney, MD, MPH
David R. Jacobs, Jr, PhD
Kiang Liu, PhD
N
UMEROUS CLINICAL INVESTI-
gations have established a
strong association between
low cardiorespiratory fitness and mortality.1-7 Cardiovascular
diseases (CVDs) account for a large proportion of mortality in adults older than
45 years.8 Numerous risk factors for
CVD, including hypertension, diabetes, and hypercholesterolemia, are suspected to be influenced by fitness,9-13
and these factors may mediate the association between low fitness and mortality. However, this proposed association is complex because fitness modifies
body mass, which is also implicated in
the development of CVD risk factors.
While previous research acknowledges the contribution of these risk
factors on mortality, relatively few studies have investigated the role of objectively determined fitness on the development of CVD risk factors in healthy
adults.
We investigated whether low cardiorespiratory fitness in young adults
(hereafter referred to as “fitness”),
estimated by shorter duration on an
exercise treadmill test, was associated
with the development of CVD risk factors (ie, hypertension, diabetes, the
Context Low cardiorespiratory fitness is an established risk factor for cardiovascular
and total mortality; however, mechanisms responsible for these associations are uncertain.
Objective To test whether low fitness, estimated by short duration on a maximal
treadmill test, predicted the development of cardiovascular disease risk factors and
whether improving fitness (increase in treadmill test duration between examinations)
was associated with risk reduction.
Design, Setting, and Participants Population-based longitudinal cohort study of
men and women 18 to 30 years of age in the Coronary Artery Risk Development in
Young Adults (CARDIA) study. Participants who completed the treadmill examination according to the Balke protocol at baseline were followed up from 1985-1986 to
2000-2001. A subset of participants (n = 2478) repeated the exercise test in 19921993.
Main Outcome Measures Incident type 2 diabetes, hypertension, the metabolic
syndrome (defined according to National Cholesterol Education Program Adult Treatment Panel III), and hypercholesterolemia (low-density lipoprotein cholesterol ⱖ160
mg/dL [4.14 mmol/L]).
Results During the 15-year study period, the rates of incident diabetes, hypertension, the metabolic syndrome, and hypercholesterolemia were 2.8, 13.0, 10.2, and
11.7 per 1000 person-years, respectively. After adjustment for age, race, sex, smoking, and family history of diabetes, hypertension, or premature myocardial infarction,
participants with low fitness (⬍20th percentile) were 3- to 6-fold more likely to develop diabetes, hypertension, and the metabolic syndrome than participants with high
fitness (ⱖ60th percentile), all P⬍.001. Adjusting for baseline body mass index diminished the strength of these associations to 2-fold (all P⬍.001). In contrast, the association between low fitness and hypercholesterolemia was modest (hazard ratio [HR],
1.4; 95% confidence interval [CI], 1.1-1.7; P=.02) and attenuated to marginal significance after body mass index adjustment (P=.13). Improved fitness over 7 years
was associated with a reduced risk of developing diabetes (HR, 0.4; 95% CI, 0.2-1.0;
P=.04) and the metabolic syndrome (HR, 0.5; 95% CI, 0.3-0.7; P⬍.001), but the strength
and significance of these associations was reduced after accounting for changes in weight.
Conclusions Poor fitness in young adults is associated with the development of cardiovascular disease risk factors. These associations involve obesity and may be modified by improving fitness.
www.jama.com
JAMA. 2003;290:3092-3100
Author Affiliations: Department of Preventive Medicine, Feinberg School of Medicine, Northwestern University, Chicago, Ill (Drs Carnethon and Liu); Nemours
Cardiac Center and Department of Pediatrics, Thomas
Jefferson University, Wilmington, Del (Drs Gidding and
Nehgme); Division of Research, Kaiser Permanente
Medical Care Program, Oakland, Calif (Dr Sidney); and
3092 JAMA, December 17, 2003—Vol 290, No. 23 (Reprinted)
Division of Epidemiology, University of Minnesota
School of Public Health, Minneapolis (Dr Jacobs).
Corresponding Author: Mercedes R. Carnethon, PhD,
Department of Preventive Medicine, Northwestern
University, 680 N Lake Shore Drive, Suite 1102, Chicago, IL 60611 (e-mail: carnethon@northwestern
.edu).
©2003 American Medical Association. All rights reserved.
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007
FITNESS AND DEVELOPMENT OF CARDIOVASCULAR RISK FACTORS
metabolic syndrome, hypercholesterolemia) independently or in association with obesity. Among participants
who repeated the exercise test 7 years
after baseline, we examined whether
improving fitness was associated with
a risk reduction.
METHODS
Study Population
The Coronary Artery Risk Development in Young Adults (CARDIA)
study is a multicenter longitudinal
cohort study designed to investigate
the development of coronary heart
disease risk factors in young adults.
Black and white men and women aged
18 to 30 years from 4 geographic areas
(Birmingham, Ala; Chicago, Ill; Minneapolis, Minn; and Oakland, Calif )
were recruited and examined in 19851986 (n = 5115).14 Participants were
reexamined at years 2 (n = 3800
[84.7%]), 5 (n = 4177 [93.1%]), 7
(n = 3922 [87.4%]), 10 (n = 3804
[84.8%]), and 15 (n = 3550 [79.1%]).
The majority of participants returned
to all 5 follow-up examinations
(n = 2728 [60.8%]), while 935
(20.8%), 445 (9.9%), 159 (3.5%), and
220 (4.9%) returned to 4, 3, 2, and 1
examination, respectively. At year 15,
there were 132 verified deaths in this
cohort. The top 3 causes of mortality
were AIDS (32%), homicide (17%),
and unintentional injury (11%), with
only 3% due to coronary heart disease
(3 definite coronary heart disease and
1 definite stroke).
We excluded participants who did
not complete the exercise treadmill test
at baseline (n=214),15 women who were
pregnant at baseline (n = 3), participants who used ␤-blockers at baseline
(n=2), and participants who did not return to follow-up examinations after
year 2 (n=409). In this cohort of 4487,
participants with a given prevalent condition under study were excluded from
incident analyses of that risk factor.
Analyses of incident hypertension, diabetes, the metabolic syndrome, and
hypercholesterolemia included 4392,
4464, 4272, and 4126 participants,
respectively. The study was approved
by the institutional review board and
written informed consent was obtained from all participants.
Data Collection
Fitness Assessment. All medically eligible participants underwent a graded
symptom-limited maximal exercise test
according to a modified Balke protocol.15 Participants were ineligible for the
following reasons: history of ischemic
heart disease, use of cardiovascular
medications other than antihypertensives, blood pressure (BP) greater than
160/90 mm Hg, or febrile illness. The
test included up to nine 2-minute stages
of progressively increasing difficulty.
Participants were encouraged to exercise as long as possible to maximal exertion. At the end of each stage, heart
rate and BP were measured.
Fitness was determined based on the
duration of the treadmill test. Participants in the lowest quintile (women: 0.2
to ⱕ6.4 minutes; men: 1.9 to ⱕ10 minutes) of sex-specific duration of testing were classified as in the low fitness
category, participants in the 20th to
60th percentile (women: 6.4-8.6 minutes; men: 10-12 minutes) were classified as moderately fit, and participants above the 60th percentile of test
duration were classified as highly fit.
These cutpoints were selected to emphasize the group at highest risk
(⬍20th percentile) and have been used
in previous studies of the association
between fitness and mortality.5,16 Resting heart rate was measured in the
seated position. Heart rate increase from
resting to maximum was calculated as
the difference between the standing
heart rate prior to the treadmill test and
the maximum heart rate during the test.
At each stage, participants were asked
to rate their level of exertion on the Borg
scale.17 Metabolic equivalent (MET) rate
of energy expenditure at peak exercise (equal to a multiple of oxygen consumption of 3.5 mL⫻kg−1 ⫻min−1) was
estimated.
Fitness testing was conducted according to the same protocol at year 7
(1992-1993) in 2478 participants from
the Chicago, Birmingham, and Oak-
©2003 American Medical Association. All rights reserved.
land study sites (participants from Minnesota were not included because of a
protocol violation).18 In this subset of
participants, changes in fitness were calculated as the difference in test duration between the year 7 and baseline
examinations. Fitness change was categorized by approximate quintiles in the
full sample, the lowest 20% decreased
duration (−12 to −2.4 minutes), the
middle 20th to 80th percentile (−2.4 to
0.4 minutes) remained stable, and the
uppermost 20% increased duration
(0.4-8.9 minutes).
Other Measurements. Participants
were asked to fast for at least 12 hours
prior to examination and avoid smoking or engaging in heavy physical activity for at least 2 hours before the
examination. Laboratory measurements of glucose and lipids19,20 were
collected according to standardized
CARDIA procedures14 and processed at
central laboratories. After 5 minutes of
rest, BP was measured from participants in the seated position 3 times at
1-minute intervals; the average of the
last 2 measurements was used. Age,
race, education, cigarette smoking status, medication use, and first-degree
family history of hypertension, diabetes, and premature (age ⬍60 years)
myocardial infarction were ascertained by interview. Height, weight, and
waist circumference (the average of 2
measurements at the minimum waist
girth) were measured on participants
in light examination clothes and no
shoes. Body mass index (BMI) was calculated as the ratio of weight (kilograms) to standing height (meters)
squared (kg/m2); obesity was defined as
BMI of at least 30. Self-reported physical activity was assessed by a validated
interview–administered questionnaire; activity scores were computed by
multiplying the frequency of participation by intensity of the activity.21
CVD Risk Factors. The incidence of
each CVD risk factor was identified as
a diagnosis of that condition at any follow-up examination among the sample
free from that condition at baseline. Diabetes was defined as a fasting glucose
level of at least 126 mg/dL (7.0 mmol/L)
(Reprinted) JAMA, December 17, 2003—Vol 290, No. 23
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007
3093
FITNESS AND DEVELOPMENT OF CARDIOVASCULAR RISK FACTORS
at examinations 7, 10, or 15 (glucose
was not assayed in years 2 and 5) or the
self-reported use of oral hypoglycemic medications or insulin at any examination. Hypertension was defined
as systolic BP of at least 140 mm Hg,
diastolic BP of at least 90 mm Hg, or
antihypertensive medication use.22
Participants were classified as having the metabolic syndrome according to the National Cholesterol Education Program/Adult Treatment Panel III
definition of at least 3 of the following: glucose level at least 110 mg/dL (6.1
mmol/L); high BP (systolic BP ⱖ130 or
diastolic BP ⱖ85 mm Hg); waist circumference greater than 88 cm (women) or greater than 102 cm (men); triglyceride level at least 150 mg/dL (1.70
mmol/L); high-density lipoprotein
(HDL) cholesterol level less than 50
mg/dL (1.30 mmol/L) in women or less
than 40 mg/dL (1.04 mmol/L) in men.23
Participants who reported using diabetes or hypertension control medications were classified as having high glucose or high BP components of the
metabolic syndrome. Hypercholesterolemia was defined as low-density lipoprotein (LDL) level of at least 160
mg/dL (4.14 mmol/L) at any follow-up examination.23
Statistical Analysis
Baseline characteristics of the population were estimated by fitness category
for men and women. Trends in covariates by fitness (ie, test duration) were estimated using F tests. Person-time was
calculated from the baseline examination until each risk factor developed or
until the last examination, whichever
came first. Incidence rates (per 1000 person-years) were calculated using Poisson regression. The proportional hazards assumption was confirmed using
log-log survival plots. We used Cox proportional hazards regression to evaluate the risk of developing each risk factor by fitness category (using 2 indicator
variables) and per-minute decrement in
test duration.
We calculated the proportion of CVD
morbidity in the population attributable to low fitness (PAR) with the fol-
lowing formula: Pe ⫻ [1 – (1/HRadj)],
where Pe is the proportion of persons
with the CVD risk factor (eg, hypertension) who are in the low fitness category and HRadj is the multivariable adjusted hazards ratio (HR) for low fitness
and that CVD risk factor.24
Next, we tested the association between fitness change over 7 years and
the development of CVD risk factors using proportional hazards regression. To
evaluate whether the associations were
similar across race, sex, and baseline
obesity, we included multiplicative interaction terms between each covariate of interest and fitness and conducted stratified analyses. The presence
of interaction was determined by the
significance of the interaction term and
stratum-specific estimates that differed markedly from each other. Because results were consistent across race
and sex, we present pooled analyses.
However, due to evidence of effect
modification by baseline obesity, we
present additional estimates specific to
strata of baseline obesity.
In a secondary analysis, we performed a propensity analysis25 in which
we used a logistic regression model to
predict high vs moderate or low fitness. The following variables were selected to calculate the score using forward stepwise modeling (P⬍.05): age,
race, sex, education, BMI, cigarette
smoking, family history of premature
myocardial infarction, self-reported
physical activity, waist circumference,
and total cholesterol level. We then
evaluated the association between fitness and each CVD risk factor adjusted for the predicted propensity score
in a proportional hazards model. Statistical significance for all analyses was
denoted at P⬍.05. All analyses were
conducted using the SAS statistical software, version 8.02 (SAS Inc, Cary, NC).
RESULTS
At baseline, less fit women (TABLE 1)
and men (TABLE 2) were slightly older
than those in the moderate and high fitness categories, had a lower education
level, and CVD risk factors were often
less favorable. Exercise test param-
3094 JAMA, December 17, 2003—Vol 290, No. 23 (Reprinted)
eters also varied by fitness category
(TABLE 3). Most associations were consistent with expectations and similar
across sex.
Hypertension, diabetes, the metabolic syndrome, and hypercholesterolemia developed in 648 (rate=13/1000
person-years), 156 (rate = 2.8), 556
(rate=10.2), and 477 (rate=11.7) persons, respectively. Low and moderate
fitness were strongly (3- to 6-fold risk
elevation) associated with the development of hypertension, diabetes, and
the metabolic syndrome and modestly
associated with an increased risk for developing hypercholesterolemia (40% increase) (FIGURE 1). Following adjustment for baseline BMI, low and
moderate fitness remained associated
with a doubling in risk for developing
each CVD risk factor except hypercholesterolemia, which attenuated to nonsignificance (low vs high, P=.13).
The number of persons in the low fitness category who developed hypertension, diabetes, the metabolic syndrome, and hypercholesterolemia was
244 (38%), 74 (47%), 202 (36%), and
99 (21%), respectively, and the adjusted PAR of developing each risk factor due to low fitness was 21%, 28%,
21%, and 4%. There was a linear relation between fitness and CVD risk factors that remained significant following adjustment for baseline BMI. The
HRs of developing hypertension, diabetes, the metabolic syndrome, and hypercholesterolemia were 1.19 (95% CI,
1.14-1.24; P⬍.001), 1.12 (95% CI, 1.031.22; P = .01), 1.16 (95% CI, 1.111.21; P⬍.001), and 1.07 (95% CI, 1.021.12; P=.01) per-minute decrement in
test duration, respectively. When we adjusted additionally for weight change
over follow-up, the magnitude and significance of the observed associations
did not change.
The distribution of fitness varied by
obesity status. Mean test duration was
10.2 minutes (95% CI, 10.1-10.3 minutes) among participants who were not
obese and 6.7 minutes (95% CI, 6.56.9 minutes) among obese participants
(P⬍.001). Only 518 (13%) nonobese
participants were in the low fitness cat-
©2003 American Medical Association. All rights reserved.
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007
FITNESS AND DEVELOPMENT OF CARDIOVASCULAR RISK FACTORS
egory compared with 352 (68%) obese
participants. Similar differences were observed for moderate (1425 [36%] vs 148
[29%], not obese vs obese), and highly
fit (2022 [51%] vs 19 [4%], not obese
vs obese) participants (all P⬍.001). Obesity emerged as an effect modifier of the
association between fitness and the development of diabetes and the metabolic syndrome. Low fitness was only as-
sociated with the development of
diabetes and the metabolic syndrome
among participants who were not already obese at baseline (TABLE 4). Patterns were similar to those in the total
population, including evidence of a dose
response.
In the subset of 2478 participants
who repeated the treadmill test, duration decreased, on average, 1 minute
(95% CI, −1.1 to −0.9 minute) between examinations. As expected due
to regression to the mean, participants in the higher fitness category at
baseline experienced significantly
greater declines in test duration (−1.4
minutes; 95% CI, −1.5 to –1.3) compared with participants in the moderate category (−0.9 minutes; 95% CI,
−1.0 to −0.8), or low fitness category
Table 1. Baseline Characteristics of the Study Population by Fitness Category Among Women (n = 2458)*
Fitness Category (Time on Treadmill)
Demographics
Age, mean (95% CI), y
Total
Low (⬍6.4 min)
(n = 493)
Moderate
(6.4-8.6 min)
(n = 956)
High
(⬎8.6 min)
(n = 1009)
P
Value
24.9 (24.8-25.1)
25.5 (25.2-25.9)
24.5 (24.2-24.7)
25.1 (24.8-25.3)
⬍.001
1270 (51.7)
409 (83.0)
580 (60.7)
281 (27.9)
⬍.001
Less than high school education, No. (%)
166 (6.8)
47 (9.6)
84 (8.8)
35 (3.5)
⬍.001
Cigarette smoking, No. (%)
Current
689 (28.2)
175 (35.7)
302 (31.7)
212 (21.2)
⬍.001
339 (13.9)
58 (11.8)
116 (12.2)
165 (16.5)
Black race, No. (%)
Former
Self-reported total physical activity,
mean (95% CI), exercise units
Family history, No. (%)
Diabetes
Hypertension
Premature myocardial infarction (age ⬍60 y)
Physical examination
Body mass index, mean (95% CI)†
336.3 (326.5-346.2)
237.9 (220.8-255.0)
290.0 (276.1-303.8)
428.8 (412.1-445.5)
.01
⬍.001
761 (31.0)
221 (44.8)
308 (32.3)
232 (23.0)
⬍.001
1780 (72.5)
393 (79.7)
685 (71.7)
702 (69.7)
⬍.001
442 (18.0)
126 (25.6)
180 (18.9)
136 (13.5)
⬍.001
24.5 (24.3-24.8)
30.6 (30.0-31.3)
24.3 (24.0-24.6)
21.8 (21.6-21.9)
⬍.001
Waist circumference, mean (95% CI), cm
74.2 (73.8-74.7)
85.6 (84.3-87.0)
74.1 (73.5-74.7)
68.8 (68.5-69.2)
⬍.001
Waist ⬎88 cm, No. (%)‡
273 (11.2)
187 (38.2)
Lipoprotein levels
Total cholesterol, mean (95% CI), mg/dL
LDL-C, mean (95% CI), mg/dL§
80 (8.4)
6 (0.6)
⬍.001
177.2 (175.9-178.5)
182.3 (179.3-185.4)
178.2 (176.1-180.3)
173.9 (172.0-175.8)
⬍.001
108.4 (107.2-109.6)
116.3 (113.4-119.1)
110.5 (108.5-112.4)
102.7 (100.9-104.4)
⬍.001
LDL-C ⬎160 mg/dL, No. (%)
143 (5.9)
HDL-C, mean (95% CI), mg/dL
55.6 (55.1-56.1)
44 (9.1)
61 (6.5)
51.2 (50.1-52.3)
54.5 (53.7-55.3)
38 (3.8)
⬍.001
58.8 (58.0-59.6)
⬍.001
HDL-C ⬍50 mg/dL, No. (%)‡
757 (31.8)
207 (43.9)
337 (36.6)
213 (21.5)
⬍.001
Triglycerides, mean (95% CI), mg/dL
66.2 (64.8-67.6)
73.9 (70.2-77.7)
66.7 (64.2-69.3)
62.0 (60.2-63.8)
⬍.001
Triglycerides ⬎150 mg/dL, No. (%) ‡
Blood pressure
Systolic, mean (95% CI), mm Hg
Diastolic, mean (95% CI), mm Hg
Elevated blood pressure ⬎130/85 mm Hg
or medications, No. (%)‡
Hypertension, No. (%)㛳
67 (2.8)
26 (2.8)
18 (1.8)
⬍.001
106.4 (106.1-106.8)
109.7 (108.7-110.6)
106.7 (106.1-107.2)
104.7 (104.1-105.2)
⬍.001
66.8 (66.4-67.1)
68.6 (67.7-69.4)
66.6 (66.0-67.1)
66.0 (65.5-66.5)
⬍.001
93 (3.8)
34 (1.4)
Glucose, mean (95% CI), mg/dL
23 (4.9)
80.5 (80.0-81.0)
37 (7.5)
27 (2.8)
17 (3.5)
8 (0.8)
82.4 (80.9-83.9)
80.2 (79.5-80.8)
29 (2.9)
⬍.001
9 (0.9)
⬍.001
80.0 (79.3-80.7)
Glucose ⬎110 mg/dL or medications, No. (%)‡
25 (1.1)
9 (1.9)
11 (1.2)
5 (0.5)
Diabetes, No. (%)
13 (0.6)
5 (1.1)
7 (0.8)
1 (0.1)
The metabolic syndrome, No. (%)
42 (1.8)
27 (5.7)
15 (1.6)
0
⬍.001
.04
.03
⬍.001
Abbreviations: CI, confidence interval; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.
SI conversions: To convert total cholesterol, LDL-C, and HDL-C values to mmol/L, multiply by 0.0259; to convert triglyceride values to mmol/L, multiply by 0.0113; to convert
glucose values to mmol/L, multiply by 0.0555.
*For some covariates there are missing values, so some percentages do not total to 100%.
†Calculated as weight in kilograms divided by the square of height in meters.
‡Metabolic syndrome component.
§LDL-C calculated using the Friedwald equation in 2453 women with triglyceride levels less than 400 mg/dL.
㛳Systolic blood pressure greater than 140 mm Hg or diastolic blood pressure greater than 90 mm Hg or medication use.
©2003 American Medical Association. All rights reserved.
(Reprinted) JAMA, December 17, 2003—Vol 290, No. 23
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007
3095
FITNESS AND DEVELOPMENT OF CARDIOVASCULAR RISK FACTORS
(−0.1 minutes; 95% CI, −0.3 to 0). A
detailed description is published elsewhere.18
The rates of hypertension, diabetes,
the metabolic syndrome, and hypercholesterolemia in this subpopulation
were 12.0 (n events=362), 2.6 (n=86),
10.0 (n=306), and 12.5 (n = 293) per
1000 person-years, respectively. After
adjustment for age, sex, race, smoking
status, and family history of diabetes,
improving fitness (increasing duration) was associated with an approximate 50% risk reduction for developing diabetes and the metabolic
syndrome, but was not associated with
the incidence of hypertension or hypercholesterolemia (FIGURE 2). Each
minute of increase in treadmill test duration was associated with nonsignifi-
cant reductions in the HRs for developing hypertension (HR, 0.99; 95% CI,
0.93-1.04; P=.63) and hypercholesterolemia (HR, 0.99; 95% CI, 0.93-1.05;
P=.70), a modest reduction in the risk
of developing diabetes (HR, 0.89; 95%
CI, 0.80-1.00; P = .05), and a significant reduction in the metabolic syndrome (HR, 0.88; 95% CI, 0.83-0.93;
P⬍.001).
Table 2. Baseline Characteristics of the Study Population by Fitness Category Among Men (n = 2029)*
Fitness Category (Time on Treadmill)
Total
Low (⬍10 min)
(n = 377)
Moderate
(10-12 min)
(n = 618)
High (⬎12 min)
(n = 1034)
24.9 (24.7-25.0)
25.1 (24.7-25.5)
24.7 (24.4-25.0)
24.9 (24.6-25.1)
967 (47.7)
234 (62.1)
331 (53.6)
402 (38.9)
⬍.001
Less than high school education, No. (%)
174 (8.6)
50 (13.3)
65 (10.6)
59 (5.7)
⬍.001
Cigarette smoking, No. (%)
Current
614 (30.5)
153 (40.9)
209 (34.2)
252 (24.5)
⬍.001
251 (12.5)
37 (9.9)
71 (11.6)
143 (13.9)
Demographics
Age, mean (95% CI), y
Black race, No. (%)
Former
Self-reported total physical activity,
mean (95% CI), exercise units
Family history, No. (%)
Diabetes
Hypertension
Premature myocardial infarction (age ⬍60 y)
Physical examination
Body mass index, mean (95% CI)†
Waist circumference, mean (95% CI), cm
Waist ⬎102 cm, No. (%)‡
519.9 (506.0-533.9)
LDL-C, mean (95% CI), mg/dL§
501.6 (476.0-527.2)
577.6 (557.9-597.2)
.22
.10
⬍.001
534 (26.3)
115 (30.5)
185 (29.9)
234 (22.6)
⬍.001
1322 (65.2)
269 (71.4)
395 (63.9)
658 (63.6)
.02
326 (16.1)
69 (18.3)
103 (16.7)
154 (14.9)
.27
24.4 (24.2-24.6)
27.4 (26.8-27.9)
24.7 (24.4-24.9)
23.2 (23.0-23.3)
⬍.001
81.8 (81.4-82.2)
88.8 (87.5-90.0)
82.3 (81.6-83.0)
78.9 (78.5-79.3)
⬍.001
71 (3.5)
Lipoprotein levels
Total cholesterol, mean (95% CI), mg/dL
391.8 (364.8-418.9)
P for
Trend
55 (14.6)
14 (2.3)
2 (0.2)
⬍.001
176.5 (175.0-178.0)
184.3 (180.6-188.0)
180.4 (177.4-183.4)
171.3 (169.5-173.2)
⬍.001
110.3 (108.9-111.7)
116.6 (113.1-120.1)
114.5 (111.8-117.2)
105.6 (103.9-107.3)
⬍.001
LDL-C ⬎160 mg/dL, No. (%)
144 (7.2)
HDL-C, mean (95% CI), mg/dL
50.1 (49.6-50.7)
39 (10.6)
48.0 (46.6-49.4)
65 (10.7)
48.8 (47.8-49.7)
40 (3.9)
⬍.001
51.7 (51.0-52.5)
⬍.001
HDL-C ⬍40 mg/dL, No. (%)‡
343 (17.5)
103 (28.9)
117 (19.7)
123 (12.2)
⬍.001
Triglycerides, mean (95% CI), mg/dL
80.5 (77.8-83.1)
99.8 (92.1-107.5)
87.1 (81.2-92.9)
69.7 (67.2-72.3)
⬍.001
Triglycerides ⬎150 mg/dL, No. (%)‡
156 (8.0)
Blood pressure
Systolic, mean (95% CI), mm Hg
58 (16.3)
62 (10.4)
36 (3.6)
⬍.001
114.9 (114.5-115.4)
117.2 (116.1-118.3)
115.1 (114.4-115.9)
113.9 (113.3-114.6)
⬍.001
Diastolic, mean (95% CI), mm Hg
70.6 (70.1-71.0)
72.2 (71.2-73.3)
70.5 (69.7-71.3)
70.0 (69.4-70.5)
⬍.001
Elevated blood pressure ⬎130/85 mm Hg
or medications, No. (%)‡
247 (12.2)
102 (9.9)
⬍.001
18 (1.7)
⬍.001
Hypertension, No. (%)㛳
60 (3.0)
Glucose, mean (95% CI), mg/dL
84.4 (83.8-85.0)
Glucose ⬎110 mg/dL or medications, No. (%)‡
22 (1.1)
71 (18.9)
23 (6.1)
19 (3.1)
85.2 (83.6-86.8)
6 (1.7)
74 (12.0)
85.2 (83.7-86.7)
7 (1.2)
83.6 (83.0-84.1)
9 (0.9)
.03
.47
Diabetes, No. (%)
10 (0.5)
4 (1.1)
2 (0.3)
4 (0.4)
.20
The metabolic syndrome, No. (%)
46 (2.3)
29 (8.1)
14 (2.4)
3 (0.3)
⬍.001
Abbreviations: CI, confidence interval; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.
SI conversions: To convert total cholesterol, LDL-C, and HDL-C values to mmol/L, multiply by 0.0259; to convert triglyceride values to mmol/L, multiply by 0.0113; to convert
glucose values to mmol/L, multiply by 0.0555.
*For some covariates there are missing values, so some percentages do not total to 100%.
†Calculated as weight in kilograms divided by the square of height in meters.
‡Metabolic syndrome component.
§LDL-C calculated using the Friedwald equation in 2018 men with triglyceride levels less than 400 mg/dL.
㛳Systolic blood pressure greater than 140 mm Hg or diastolic blood pressure greater than 90 mm Hg or medication use.
3096 JAMA, December 17, 2003—Vol 290, No. 23 (Reprinted)
©2003 American Medical Association. All rights reserved.
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007
FITNESS AND DEVELOPMENT OF CARDIOVASCULAR RISK FACTORS
15 years (r=−0.25) (both P ⬍.001). Following adjustment for baseline BMI and
weight change over 15 years, the HR attenuated to marginal significance for diabetes (HR, 0.89; 95% CI, 0.78-1.01 per
1-minute increase; P=.06) and the metabolic syndrome (HR, 0.96; 95% CI, 0.901.02 per 1-minute increase; P=.20).
In the subset of participants who performed the treadmill test at baseline and
7 years after baseline, participants
gained, on average, 7 kg (95% CI, 6.77.2 kg) over 7 years and 12.7 kg (95%
CI, 12.3-13.1 kg) over 15 years. Change
in fitness was inversely correlated with
weight gain over 7 years (r=−0.37) and
In the total study population
(n=4487), propensity scores were generated from demographic, behavioral,
and clinical measurements. Scores
ranged from 0 to 0.98. Propensity score–
adjusted HRs for developing incident
CVD risk factors (TABLE 5) did not differ appreciably from multivariable ad-
Table 3. Baseline Characteristics of the Exercise Test by Fitness Among Women and Men*
Fitness Category†
Total
Low
Women (n = 2458)
Moderate
High
Exercise test duration, min
Resting heart rate, beats/min
Heart rate at stage 2 (4 min), beats/min
8.3 (8.2-8.3)
71.8 (71.4-72.3)
137.6 (136.8-138.3)
5.2 (5.1-5.3)
72.7 (71.7-73.6)
151.6 (150.0-153.2)
7.6 (7.5-7.6)
73.1 (72.4-73.7)
140.4 (139.3-141.4)
10.4 (10.3-10.5)
70.3 (69.7-70.9)
128.2 (127.3-129.1)
Maximum heart rate, beats/min
Heart rate increase from resting
to maximum, beats/min
Maximum rating of perceived
exertion (range, 6-20)
Estimated METs
177.1 (176.5-177.7)
94.9 (94.2-95.6)
165.9 (164.3-167.4)
81.5 (79.9-83.1)
176.8 (175.9-177.6)
92.9 (91.9-93.9)
183.0 (182.3-183.7)
103.3 (102.4-104.2)
17.8 (17.7-17.8)
17.0 (16.8-17.2)
17.6 (17.5-17.8)
18.3 (18.2-18.4)
10.6 (10.5-10.6)
7.7 (7.6-7.8)
9.9 (9.8-9.9)
12.6 (12.5-12.7)
Men (n = 2029)
11.7 (11.6-11.8)
8.3 (8.2-8.4)
65.9 (65.5-66.4)
69.4 (68.4-70.5)
114.7 (114.0-115.4)
124.6 (122.9-126.3)
180.8 (180.1-181.5)
167.8 (165.8-169.9)
108.0 (107.2-108.8)
92.2 (90.3-94.2)
Exercise test duration, min
Resting heart rate, beats/min
Heart rate at stage 2 (4 min), beats/min
Maximum heart rate, beats/min
Heart rate increase from resting
to maximum, beats/min
Maximum rating of perceived
exertion (range, 6-20)
Estimated METs
10.9 (10.8-10.9)
66.6 (65.8-67.3)
116.8 (115.5-118.0)
178.7 (177.6-179.8)
104.8 (103.6-105.9)
13.5 (13.4-13.5)
64.3 (63.7-64.8)
109.9 (109.1-110.7)
186.8 (186.1-187.6)
115.7 (114.7-116.6)
18.1 (18.0-18.2)
17.6 (17.3-17.8)
18.1 (18.0-18.3)
18.3 (18.2-18.4)
13.7 (13.6-13.8)
10.7 (10.5-10.8)
13.0 (12.9-13.1)
15.2 (15.1-15.3)
Abbreviation: MET, metabolic equivalent.
*Data are presented as means with 95% confidence intervals. P for trend, ⬍.001.
†Women: low, less than 6.4 minutes; moderate, 6.4-8.6 minutes; high, more than 8.6 minutes. Men: low, less than 10 minutes; moderate, 10-12 minutes; high, more than 12
minutes.
Figure 1. Adjusted Hazard Ratios of the Association Between Fitness Category and Incident Cardiovascular Disease Risk Factors
Model 1
Model 2
10.0
Hazard Ratios
(95% Confidence Intervals)
Hazard Ratios
(95% Confidence Intervals)
10.0
1.0
0.1
Hypertension
Diabetes
The
Metabolic
Syndrome
Hypercholesterolemia∗
Risk Factors
Fitness Category
Low
Moderate
1.0
0.1
Hypertension
Diabetes
The
Metabolic
Syndrome
Hypercholesterolemia∗
Risk Factors
Comparison for both models is high fitness category. Model 1: Adjusted for age, race, sex, current smoking, family history of hypertension (hypertension) or family
history of diabetes (diabetes and the metabolic syndrome) or family history of myocardial infarction younger than 60 years of age (hypercholesterolemia). Model 2:
Adjusted as for model 1 plus baseline body mass index.
*Incident low-density lipoprotein cholesterol ⱖ160 mg/dL (4.14 mmol/L).
©2003 American Medical Association. All rights reserved.
(Reprinted) JAMA, December 17, 2003—Vol 290, No. 23
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007
3097
FITNESS AND DEVELOPMENT OF CARDIOVASCULAR RISK FACTORS
justed results. The risk of developing hypertension, diabetes, and the metabolic
syndrome remained double among participants in the low fitness category compared with the high fitness category.
COMMENT
Cardiorespiratory fitness in young men
and women, estimated by the duration of a maximal treadmill exercise test,
was inversely associated with the risk
of developing hypertension, diabetes,
metabolic syndrome, and hypercholesterolemia in middle age. Our findings
were only partly attributable to body
mass and weight maintenance, suggest-
Table 4. Adjusted Hazard Ratios of Baseline Fitness Categories and Incident Cardiovascular Disease Risk Factors, Stratified by Baseline
Obesity*
Nonobese (BMI ⬍30)†
Hypertension
Low
Moderate
High
Per 1-min decrement
Diabetes
Low
Moderate
High
Per 1-min decrement
The metabolic syndrome
Low
Moderate
High
Per 1-min decrement
Hypercholesterolemia
Low
Moderate
High
Per 1-min decrement
Obese (BMI ⱖ30)†
No.
No. of
Events
Rate‡
Hazard Ratio
(95% CI)
No.
No. of
Events
Rate‡
Hazard Ratio
(95% CI)
P
Interaction
501
1400
1996
3897
119
199
161
479
19.0
10.9
5.9
9.3§
2.59 (2.02-3.32)
1.80 (1.46-2.23)
1.00
1.21 (1.16-1.26)
328
146
18
492
125
38
6
169
32.2
20.5
27.0
28.4§
1.27 (0.56-2.88)
0.85 (0.36-2.02)
1.00
1.23 (1.14-1.32)
.10
.10
513
1419
2017
3949
23
35
24
82
3.4
1.9
0.9
1.5§
3.66 (2.02-6.63)
2.18 (1.28-3.69)
1.00
1.26 (1.14-1.38)
348
145
19
438
51
20
3
74
11.2
10.3
11.7
11.0§
1.06 (0.33-3.42)
1.00 (0.29-3.40)
1.00
1.03 (0.92-1.16)
.06
.25
492
1372
1982
3846
82
180
113
375
13.2
10.2
4.2
7.4§
4.05 (3.01-5.46)
2.93 (2.30-3.73)
1.00
1.28 (1.22-1.34)
285
120
18
423
120
52
9
181
35.5
37.8
40.2
36.3§
1.26 (0.63-2.49)
1.35 (0.66-2.76)
1.00
1.07 (1.00-1.14)
⬍.001
.04
466
1291
1926
3683
59
149
207
415
10.0
9.2
8.2
8.7§
1.29 (0.95-1.75)
1.24 (1.00-1.54)
1.00
1.08 (1.03-1.13)
300
122
18
440
40
19
3
62
10.6
12.2
12.9
11.1§
1.05 (0.32-3.39)
1.25 (0.37-4.25)
1.00
0.99 (0.89-1.10)
.73
.99
.76
⬍.001
⬍.001
.11
Abbreviations: BMI, body mass index; CI, confidence interval.
*Hazard ratios are adjusted for age, race, sex, baseline smoking status, family history of hypertension (hypertension) or family history of diabetes (diabetes and the metabolic syndrome) or family history of myocardial infarction younger than 60 years of age (hypercholesterolemia).
†Calculated as weight in kilograms divided by the square of height in meters.
‡Event rate per 1000 person-years.
§Event rate per 1000 person-years for the entire strata (ie, across fitness categories).
Figure 2. Adjusted Hazard Ratios of the Association Between Change in Fitness and Incident Cardiovascular Disease Risk Factors (n =2478)
Model 1
Model 2
10.0
Hazard Ratios
(95% Confidence Intervals)
Hazard Ratios
(95% Confidence Intervals)
10.0
1.0
0.1
Hypertension
Diabetes
The
Metabolic
Syndrome
Hypercholesterolemia∗
Risk Factors
Fitness Status
Stable
Increased
1.0
0.1
Hypertension
Diabetes
The
Metabolic
Syndrome
Hypercholesterolemia∗
Risk Factors
Comparison for both models is decreased fitness. Model 1: Adjusted for age, race, sex, baseline smoking status, and family history of hypertension (hypertension) or
family history of diabetes (diabetes and metabolic syndrome) or family history of myocardial infarction younger than 60 years of age (hypercholesterolemia). Model 2:
Adjusted as for model 1 plus baseline body mass index and weight change over follow-up.
*Incident low-density lipoprotein cholesterol ⱖ160 mg/dL (4.14 mmol/L).
3098 JAMA, December 17, 2003—Vol 290, No. 23 (Reprinted)
©2003 American Medical Association. All rights reserved.
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007
FITNESS AND DEVELOPMENT OF CARDIOVASCULAR RISK FACTORS
ing that fitness plays an important independent protective role in the development of cardiovascular risk factors.
However, among those who became
obese earlier in life (possibly during
childhood or adolescence), fitness does
not protect against developing diabetes or metabolic syndrome. Increasing
treadmill test duration between visits
was associated with a lower risk for developing both diabetes and the metabolic syndrome, suggesting that 2 very
important risk factors for coronary heart
disease and mortality may be modified by improved fitness over time.
Persons who are physically fit maintain a more favorable caloric balance and
lower body weights, both of which protect against the development of CVD risk
factors. Previous observational and trial
results demonstrate that leaner persons and persons who lose weight are
at markedly lower risk for developing
hypertension,26,27 diabetes,28 the metabolic syndrome, 29 and lipid disorders.30 A recent meta-analysis of 44 studies of the effect of physical fitness on BP
reported BP declines with fitness training in both lean and obese study subjects. Mechanisms suggested to account for these observations are reduced
systemic vascular resistance, decreased
cardiac output, and decreased plasma
noradrenaline concentrations.31
Fitness promotes muscle insulin sensitivity,12 insulin-mediated transport of
glucose from blood to muscles,32 improved autonomic nervous system
function,33 and lower heart rates, which
each decrease the risk of developing diabetes, independent of body mass. Increased lipoprotein lipase activity in active skeletal muscle, which results in an
enhanced clearance rate of plasma triglycerides; increased transport of lipids and lipoproteins from the peripheral circulation and tissues to the liver;
and enhanced HDL cholesterol, are
mechanisms by which lipids may improve with fitness.10,34,30 The relatively
modest association between fitness and
high LDL cholesterol levels in this study
may be attributable to the greater contribution of genetics and diet than fitness to LDL levels.
Table 5. Propensity Score Adjusted Hazard Ratios of the Association Between
Cardiorespiratory Fitness and Cardiovascular Disease Risk Factors*
Risk Factors
Low Fitness vs High Fitness,
Hazard Ratio
(95% Confidence Interval)
Moderate Fitness vs High Fitness,
Hazard Ratio
(95% Confidence Interval)
Hypertension
Diabetes
The metabolic syndrome
2.17 (1.69-2.78)
1.75 (1.01-3.04)
1.87 (1.42-2.48)
1.34 (1.07-1.67)
1.25 (0.75-2.09)
1.64 (1.29-2.09)
Hypercholesterolemia†
1.02 (0.76-1.36)
1.01 (0.81-1.27)
*Variables included in the propensity score are age, race, sex, education (at least high school education vs greater
than high school education), body mass index, current vs never smoking, family history of premature myocardial
infarction, self-reported physical activity, waist circumference, and total cholesterol value.
†Incident low-density lipoprotein cholesterol at least 160 mg/dL (4.14 mmol/L).
Our results must be interpreted in
light of several limitations. First, in this
study we used treadmill test duration
as an estimate of fitness in lieu of direct measurements of maximum oxy·
gen consumption per unit time (VO2).
Previous research has demonstrated a
strong correlation (r = 0.92) between
test duration on a symptom-limited test
·
and VO2.35 Next, it is possible that our
measure of improving fitness, increasing duration of the exercise treadmill
test 7 years later, may simply reflect increased familiarity with the apparatus
among participants using it for at least
the second time. Finally, because physical fitness reflects the combination of
genetics and physical activity,36 fitness may not represent a truly modifiable behavior. However, the strong association between self-reported
participation in physical activity and test
duration suggests that we are at least
in part capturing the role of physical activity participation on health risk.
Our findings demonstrate the importance of low cardiorespiratory fitness in
young adulthood as a risk factor for developing cardiovascular comorbidities in
middle age. Previous work has demonstrated that engaging in a regular exercise program can improve fitness.37 If the
association between fitness and CVD risk
factor development is causal, and if all
unfit young adults had been fit, there may
have been 21% to 28% fewer cases of hypertension, diabetes, and metabolic syndrome. Given the current obesity epidemic and observations of a decline in
daily energy expenditure in the population,38 improving cardiorespiratory fitness in young men and women and de-
©2003 American Medical Association. All rights reserved.
veloping public health policies that
encourage physical activity should be important health policy goals. Substantial
public health benefits may be achieved
in the prevention of CVD morbidity by
reversing adverse trends in the general
level of fitness in the population.
Author Contributions: As principal investigator, Dr
Carnethon had full access to all the data in the study
and takes responsibility for the integrity of the data
and the accuracy of the data analysis. Study concept
and design: Carnethon, Gidding, Jacobs.
Acquisition of data: Gidding, Sidney, Jacobs, Liu.
Analysis and interpretation of data: Carnethon,
Gidding, Nehgme, Sidney, Jacobs, Liu.
Drafting of the manuscript: Carnethon, Gidding, Liu.
Critical revision of the manuscript for important intellectual content: Carnethon, Gidding, Nehgme,
Sidney, Jacobs, Liu.
Statistical expertise: Carnethon, Gidding, Jacobs, Liu.
Obtained funding: Sidney, Jacobs, Liu.
Study supervision: Gidding, Liu.
Funding/Support: Work on this article was partially
supported by contracts N01-HC-48047, N01-HC48048, N01-HC-48049, N01-HC-48050 and N01HC-95095 from the National Institutes of Health. Dr
Carnethon was supported in part by career development award 1 K01 HL73249-01 from the National
Heart, Lung, and Blood Institute.
Role of the Sponsor: The National Heart, Lung, and
Blood Institute (NHLBI) of the National Institutes of
Health was involved in the design of the study and
data collection. A project officer assigned by the NHLBI
was responsible for oversight of analyses and manuscript preparation. Final approval of the manuscript
for submission was granted by the sponsoring organization.
REFERENCES
1. Lee CD, Blair SN. Cardiorespiratory fitness and
smoking-related and total cancer mortality in men. Med
Sci Sports Exerc. 2002;34:735-739.
2. Lee CD, Blair SN. Cardiorespiratory fitness and stroke
mortality in men. Med Sci Sports Exerc. 2002;34:592595.
3. Ekelund LG, Haskell WL, Johnson JL, Whaley FS,
Criqui MH, Sheps DS. Physical fitness as a predictor
of cardiovascular mortality in asymptomatic North
American men: the Lipid Research Clinics Mortality Follow-up Study. N Engl J Med. 1988;319:1379-1384.
4. Blair SN, Kohl HW 3rd, Paffenbarger RS Jr, Clark
DG, Cooper KH, Gibbons LW. Physical fitness and allcause mortality: a prospective study of healthy men
and women. JAMA. 1989;262:2395-2401.
5. Wei M, Kampert JB, Barlow CE, et al. Relationship
(Reprinted) JAMA, December 17, 2003—Vol 290, No. 23
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007
3099
FITNESS AND DEVELOPMENT OF CARDIOVASCULAR RISK FACTORS
between low cardiorespiratory fitness and mortality
in normal-weight, overweight, and obese men. JAMA.
1999;282:1547-1553.
6. Gulati M, Pandey DK, Arnsdorf MF, et al. Exercise
capacity and the risk of death in women: the St James
Women Take Heart Project. Circulation. 2003;108:
1554-1559.
7. Mora S, Redberg RF, Cui Y, et al. Ability of exercise testing to predict cardiovascular and all-cause death
in asymptomatic women: a 20-year follow-up of the
Lipid Research Clinics Prevalence Study. JAMA. 2003;
290:1600-1607.
8. Arias E, Smith BL. Deaths: Preliminary Data for
2001. Hyattsville, Md: National Center for Health Statistics; 2003.
9. Blair SN, Goodyear NN, Gibbons LW, Cooper KH.
Physical fitness and incidence of hypertension in healthy
normotensive men and women. JAMA. 1984;252:
487-490.
10. Kraus WE, Houmard JA, Duscha BD, et al. Effects of the amount and intensity of exercise on plasma
lipoproteins. N Engl J Med. 2002;347:1483-1492.
11. Sternfeld B, Sidney S, Jacobs DR Jr, Sadler MC,
Haskell WL, Schreiner PJ. Seven-year changes in physical fitness, physical activity, and lipid profile in the
CARDIA study. Ann Epidemiol. 1999;9:25-33.
12. Wei M, Gibbons LW, Mitchell TL, Kampert JB, Lee
CD, Blair SN. The association between cardiorespiratory fitness and impaired fasting glucose and type 2
diabetes mellitus in men. Ann Intern Med. 1999;130:
89-96.
13. Laaksonen DE, Lakka H-M, Salonen JT, Niskanen LK, Rauramaa R, Lakka TA. Low levels of leisuretime physical activity and cardiorespiratory fitness predict development of the metabolic syndrome. Diabetes
Care. 2002;25:1612-1618.
14. Friedman GD, Cutter GR, Donahue RP, et al.
CARDIA: study design, recruitment, and some characteristics of the examined subjects. J Clin Epidemiol. 1988;41:1105-1116.
15. Sidney S, Haskell WL, Crow R, et al. Symptomlimited graded treadmill exercise testing in young adults
in the CARDIA study. Med Sci Sports Exerc. 1992;24:
177-183.
16. Blair SN, Kampert JB, Kohl HW 3rd, et al. Influ-
ences of cardiorespiratory fitness and other precursors on cardiovascular disease and all-cause mortality in men and women. JAMA. 1996;276:205-210.
17. Borg GA. Perceived exertion. Exerc Sport Sci Rev.
1974;2:131-153.
18. Sidney S, Sternfeld B, Haskell WL, Quesenberry
CP Jr, Crow RS, Thomas RJ. Seven-year change in
graded exercise treadmill test performance in young
adults in the CARDIA study. Med Sci Sports Exerc.
1998;30:427-433.
19. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein
cholesterol in plasma, without use of the preparative
ultracentrifuge. Clin Chem. 1972;18:499-502.
20. Warnick GR, Benderson J, Albers JJ. Dextran sulfate-Mg2+ precipitation procedure for quantitation of
high-density-lipoprotein cholesterol. Clin Chem. 1982;
28:1379-1388.
21. Jacobs DR, Hahn L, Haskell WL, Pirie P, Sidney
S. Validity and reliability of short physical activity history: CARDIA Study and the Minnesota Heart Health
Program. J Cardiopulm Rehabil. 1989;9:448-459.
22. Chobanian AV, Bakris GL, Black HR, et al. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High
Blood Pressure: the JNC 7 Report. JAMA. 2003;289:
2560-2571.
23. Executive summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert
Panel on Detection, Evalution, and Treatment of High
Blood Cholesterol in Adults (Adult Treatment Panel
III). JAMA. 2001;285:2486-2497.
24. Rothman KJ. Epidemiology: An Introduction. New
York, NY: Oxford University Press; 2002.
25. D’Agostino RB Jr. Propensity score methods for
bias reduction in the comparison of a treatment to a
non-randomized control group. Stat Med. 1998;17:
2265-2281.
26. Juhaeri, Stevens J, Chambless LE, et al. Associations between weight gain and incident hypertension in a bi-ethnic cohort: the Atherosclerosis Risk in
Communities Study. Int J Obes Relat Metab Disord.
2002;26:58-64.
27. Stamler R, Stamler J, Riedlinger WF, Algera G, Roberts RH. Weight and blood pressure: findings in hy-
pertension screening of 1 million Americans. JAMA.
1978;240:1607-1610.
28. Must A, Spadano J, Coakley EH, Field AE, Colditz
G, Dietz WH. The disease burden associated with overweight and obesity. JAMA. 1999;282:1523-1529.
29. Park YW, Zhu S, Palaniappan L, Heshka S, Carnethon MR, Heymsfield SB. The metabolic syndrome: prevalence and associated risk factor findings
in the US population from the Third National Health
and Nutrition Examination Survey, 1988-1994. Arch
Intern Med. 2003;163:427-436.
30. Durstine JL, Haskell WL. Effects of exercise training on plasma lipids and lipoproteins. Exerc Sport Sci
Rev. 1994;22:477-521.
31. Fagard RH. Physical activity in the prevention and
treatment of hypertension in the obese. Med Sci Sports
Exerc. 1999;31(11 suppl):S624-S630.
32. Perseghin G, Price TB, Petersen KF, et al. Increased glucose transport-phosphorylation and muscle
glycogen synthesis after exercise training in insulinresistant subjects. N Engl J Med. 1996;335:13571362.
33. Melanson EL, Freedson PS. The effect of endurance training on resting heart rate variability in sedentary adult males. Eur J Appl Physiol. 2001;85:442449.
34. Stefanick ML, Wood PD. Physical activity, lipid and
lipoprotein metabolism, and lipid transport. In: Stevens
T, ed. Physical Activity, Fitness and Health: International Proceedings and Consensus Statement. Champaign, Ill: Human Kinetics; 1994:417-428.
35. Pollock ML, Bohannon RL, Cooper KH, et al. A
comparative analysis of four protocols for maximal
treadmill stress testing. Am Heart J. 1976;92:39-46.
36. Physical Activity and Health: A Report of the Surgeon General. Atlanta, Ga: Centers for Disease Control and Prevention; 1996.
37. Petrella RJ, Koval JJ, Cunningham DA, Paterson
DH. Can primary care doctors prescribe exercise to improve fitness? the Step Test Exercise Prescription (STEP)
project. Am J Prev Med. 2003;24:316-322.
38. Caspersen CJ, Pereira MA, Curran KM. Changes
in physical activity patterns in the United States, by
sex and cross-sectional age. Med Sci Sports Exerc.
2000;32:1601-1609.
The wise man is satisfied with nothing.
—William Godwin (1756-1836)
3100 JAMA, December 17, 2003—Vol 290, No. 23 (Reprinted)
©2003 American Medical Association. All rights reserved.
Downloaded from www.jama.com at Medical Library of the PLA, on August 17, 2007