Content uploaded by Sara L Partington
Author content
All content in this area was uploaded by Sara L Partington
Content may be subject to copyright.
The New England
Journal
of
Medicine
Copyright © 2002 by the Massachusetts Medical Society
VOLUME 346
M
ARCH
14, 2002
NUMBER 11
N Engl J Med, Vol. 346, No. 11
·
March 14, 2002
·
www.nejm.org
·
793
EXERCISE CAPACITY AND MORTALITY AMONG MEN REFERRED
FOR EXERCISE TESTING
J
ONATHAN
M
YERS
, P
H
.D., M
ANISH
P
RAKASH
, M.D., V
ICTOR
F
ROELICHER
, M.D., D
AT
D
O
, M.D., S
ARA
P
ARTINGTON
, B.S
C
.,
AND
J. E
DWIN
A
TWOOD
, M.D.
A
BSTRACT
Background
Exercise capacity is known to be an
important prognostic factor in patients with cardiovas-
cular disease, but it is uncertain whether it predicts
mortality equally well among healthy persons. There
is also uncertainty regarding the predictive power of
exercise capacity relative to other clinical and exercise-
test variables.
Methods
We studied a total of 6213 consecutive
men referred for treadmill exercise testing for clinical
reasons during a mean (±SD) of 6.2±3.7 years of fol-
low-up. Subjects were classified into two groups: 3679
had an abnormal exercise-test result or a history of
cardiovascular disease, or both, and 2534 had a nor-
mal exercise-test result and no history of cardiovascu-
lar disease. Overall mortality was the end point.
Results
There were a total of 1256 deaths during the
follow-up period, resulting in an average annual mor-
tality of 2.6 percent. Men who died were older than
those who survived and had a lower maximal heart
rate, lower maximal systolic and diastolic blood pres-
sure, and lower exercise capacity. After adjustment for
age, the peak exercise capacity measured in metabolic
equivalents (MET) was the strongest predictor of the
risk of death among both normal subjects and those
with cardiovascular disease. Absolute peak exercise ca-
pacity was a stronger predictor of the risk of death than
the percentage of the age-predicted value achieved,
and there was no interaction between the use or non-
use of beta-blockade and the predictive power of exer-
cise capacity. Each 1-MET increase in exercise capacity
conferred a 12 percent improvement in survival.
Conclusions
Exercise capacity is a more powerful
predictor of mortality among men than other estab-
lished risk factors for cardiovascular disease. (N Engl
J Med 2002;346:793-801.)
Copyright © 2002 Massachusetts Medical Society.
From the Division of Cardiovascular Medicine, Stanford University Med-
ical Center and the Veterans Affairs Palo Alto Health Care System — both in
Palo Alto, Calif. Address reprint requests to Dr. Myers at the Cardiology
Division (111C), Veterans Affairs Palo Alto Health Care System, 3081
Miranda Ave., Palo Alto, CA 94304, or at drj993@aol.com.
URING the past two decades, exercise ca-
pacity and activity status have become well-
established predictors of cardiovascular and
overall mortality.
1,2
The fact that exercise
capacity is a strong and independent predictor of out-
comes supports the value of the exercise test as a clin-
ical tool; it is noninvasive, is relatively inexpensive, and
provides a wealth of clinically relevant diagnostic and
prognostic information.
3,4
However, recent guidelines
4
and commentaries on the topic
5,6
have identified sev-
eral areas related to the prognostic usefulness of exer-
cise testing that are in need of further study. For ex-
ample, the majority of previous studies have not clearly
assessed the independent prognostic power of exercise
capacity relative to other clinical variables and infor-
mation from exercise testing. In addition, whereas the
literature is filled with long-term follow-up studies
conducted in relatively healthy populations,
7-11
few
studies have focused on more clinically relevant pop-
ulations — that is, patients referred for exercise test-
ing for clinical reasons. Moreover, although exercise
capacity expressed in terms of metabolic equivalents
(MET) is the common clinical measure of exercise tol-
erance, exercise capacity is strongly influenced by age
and activity status. It is not known which has greater
prognostic value: the absolute peak exercise capacity
(measured in MET) or exercise capacity expressed as
a percentage of the value predicted on the basis of age.
Finally, the use of beta-blocker therapy is common
among the patients who are typically referred for exer-
cise testing; although beta-blockade improves surviv-
al, it can also reduce exercise capacity. Data related to
the influence of beta-blockade on the prognostic val-
ue of exercise tolerance are sparse.
D
794
·
N Engl J Med, Vol. 346, No. 11
·
March 14, 2002
·
www.nejm.org
The New England Journal of Medicine
In the present study, we assessed the prognostic
value of exercise capacity among patients referred for
exercise testing for clinical reasons. We addressed the
questions of whether exercise capacity is an independ-
ent predictor of the risk of death; whether it is as
strong a marker of risk as other established cardiovas-
cular risk factors; whether the percentage of age-pre-
dicted exercise capacity achieved is a better marker
of risk than the absolute peak exercise capacity; and
whether beta-blockade influences the prognostic val-
ue of exercise capacity.
METHODS
Exercise Testing
The study population consisted of 6213 consecutive men referred
for exercise testing for clinical reasons. Beginning in 1987, a thor-
ough clinical history, current medications, and risk factors in these
subjects were recorded prospectively on computerized forms at the
time of the exercise tests.
12,13
After providing written informed con-
sent, the subjects underwent symptom-limited treadmill testing ac-
cording to standardized graded
14
or individualized
15
ramp-treadmill
protocols. Before testing, the subjects were given a questionnaire,
which we used to estimate their exercise capacity; the use of this
estimate allowed most subjects to reach maximal exercise capacity
within the recommended range of 8 to 12 minutes.
16
We have pre-
viously observed that this protocol results in the closest relation
between the measured and estimated exercise capacity.
15
(One MET
is defined as the energy expended in sitting quietly, which is equiv-
alent to a body oxygen consumption of approximately 3.5 ml per
kilogram of body weight per minute for an average adult.) Subjects
were discouraged from using the handrails for support. Target heart
rates were not used as predetermined end points. Subjects were
placed in a supine position as soon as possible after exercise.
17
Med-
ications were not changed or stopped before testing.
ST-segment depression was measured visually. Ventricular tachy-
cardia was defined as a run of three or more consecutive premature
ventricular contractions, and if 10 percent or more of all ventricular
contractions were premature, the subject was considered to have
frequent premature ventricular contractions.
18
Exercise capacity (in
MET) was estimated on the basis of the speed and grade of the
treadmill.
19
Subjects with either a decrease of 10 mm Hg in sys-
tolic blood pressure after an initial increase with exercise or a de-
crease to 10 mm Hg below the value measured while standing be-
fore testing were considered to have exertional hypotension.
20
No test results were classified as indeterminate.
21
The exercise tests
were performed, analyzed, and reported according to a standardized
protocol and with the use of a computerized data base.
22
Normal
standards for age-predicted exercise capacity were derived from re-
gression equations developed on the basis of results in veterans who
were referred for exercise testing
23
and the predicted peak exercise
capacity was calculated as 18.0¡(0.15¬age). The percentage of nor-
mal exercise capacity achieved was defined as follows: (achieved ex-
ercise capacity÷the predicted energy expenditure)¬100.
We defined subjects with cardiovascular disease as those with a
history of angiographically documented coronary artery disease, my-
ocardial infarction, coronary bypass surgery, coronary angioplasty,
congestive heart failure, peripheral vascular disease, or an abnormal
result on an exercise test that was suggestive of coronary artery dis-
ease (ST-segment depression of »1.0 mm, exercise-induced angina,
or both). Seven percent of the population (435 subjects) had a his-
tory of mild pulmonary disease and were included in the group with
an abnormal exercise-test result, a history of cardiovascular disease,
or both, which included a total of 3679 subjects. The other 2534
subjects, who had no evidence of cardiovascular disease, were clas-
sified as normal.
Follow-up
The Social Security death index was used to match all subjects to
their records according to name and Social Security number. Vital
status was determined as of July 2000.
Statistical Analysis
NCSS software (Salt Lake City) was used for all statistical analy-
ses. Overall mortality was used as the end point for survival analysis.
Censoring was not performed, since data on interventions were not
available for all subjects. Survival analysis was performed with the use
of Kaplan–Meier curves for the comparison of variables and cutoff
points, and a Cox proportional-hazards model was used to deter-
mine which variables were independently and significantly associat-
ed with the time to death. Analyses were adjusted for age in single
years as a continuous variable.
In order to compare our results with those of previous studies,
the relative risk of death was calculated for each quintile of exercise
capacity; subjects with an exercise capacity of less than 5 MET were
considered to have a high risk of death, and those with an energy
expenditure of more than 8 MET were considered to have a low
risk. Receiver-operating-characteristic curves were constructed in
order to compare the absolute exercise capacity achieved and ex-
ercise capacity expressed as a percentage of the age-predicted val-
ue in terms of their discriminatory accuracy in predicting survival.
The receiver-operating-characteristic curves were compared with
the use of the z statistic.
RESULTS
The mean (±SD) follow-up period was 6.2±3.7
years, and the average annual mortality was 2.6 per-
cent. No major complications occurred, although
nonsustained ventricular tachycardia (three or more
consecutive beats) occurred during 1.1 percent of the
exercise tests. A total of 83 percent of the subjects who
were classified as normal achieved a maximal heart rate
that was at least 85 percent of the age-predicted value.
Demographic Characteristics
As compared with the normal subjects, subjects
with cardiovascular disease were older, had a slightly
lower body-mass index (defined as the weight in kilo-
grams divided by the square of the height in meters),
and had more extensive use of medicines in addition
to more cardiovascular interventions (Table 1).
Exercise-Test Results
Age-adjusted demographic characteristics and the
results of exercise testing in the subjects who sur-
vived and those who died in both groups are pre-
sented in Table 2. The regression equation that
predicted the peak exercise capacity on the basis of
age was 18.4¡(0.16¬age); with this equation,
r(±SE)=¡0.50±0.31, P<0.001. The regression
equation used to predict the maximal heart rate on
the basis of age was 187¡(0.85¬age); with this equa-
tion, r(±SE)=¡0.39±0.23, P< 0.001.
Predictors of Death from Any Cause
Clinical and exercise-test predictors of mortality
from the Cox proportional hazards model are present-
EXERCISE CAPACITY AND MORTALITY
N Engl J Med, Vol. 346, No. 11
·
March 14, 2002
·
www.nejm.org
·
795
*Plus–minus values are means ±SD. To convert values for height to centimeters, multiply by 2.54;
to convert values for weight to kilograms, multiply by 0.45. For comparisons where no P value is
given, the differences are due to the classification criteria specified in the study design.
T
ABLE
1.
D
EMOGRAPHIC
AND
C
LINICAL
C
HARACTERISTICS
OF
N
ORMAL
S
UBJECTS
AND
S
UBJECTS
WITH
C
ARDIOVASCULAR
D
ISEASE
.*
V
ARIABLE
A
LL
S
UBJECTS
(N=6213)
N
ORMAL
S
UBJECTS
(N=2534)
S
UBJECTS
WITH
C
ARDIOVASCULAR
D
ISEASE
(N=3679)
P
V
ALUE
Demographic characteristics
Age (yr) 59±11.2 55.5±11.8 61.5±10.1 <0.001
Height (in.) 69.2±4.1 69.4±3.4 69.2±3.6 0.02
Weight (lb) 191.2±39 193.7±37 188.8±36 <0.001
Body-mass index 28.0±5.2 28.4±5.1 27.3±5.0 <0.001
Medications (%)
Digoxin 5.4 0 9.1
Calcium antagonist 27.3 17.2 34.3 <0.001
Beta-blocker 18.9 12.0 23.7 <0.001
Nitrate 23.3 9.5 32.9 <0.001
Antihypertensive agent 24.0 19.3 27.3 <0.001
Medical history (%)
Atrial fibrillation 3.1 0.8 2.7 <0.001
Pulmonary disease 6.9 0 11.7
Stroke 3.6 0 6.1
Claudication 5.3 0 8.9
Typical angina 31.3 7 31.2 <0.001
Myocardial infarction 29.3 0 49.4
Congestive heart failure 8.4 0 14.2
Interventions (%)
Coronary bypass surgery 9.3 0 14.1
Percutaneous transluminal cor-
onary angioplasty, stenting,
or both
5.2 0 8.7
*Plus–minus values are means ±SD. P values are for comparisons between the subjects who survived and those who died in each group.
To convert values for height to centimeters, multiply by 2.54; to convert values for weight to kilograms, multiply by 0.45. MET denotes
metabolic equivalents.
T
ABLE
2.
A
GE
-A
DJUSTED
C
HARACTERISTICS
AND
E
XERCISE
-T
EST
R
ESPONSES
AMONG
S
UBJECTS
W
HO
D
IED
AND
S
UBJECTS
W
HO
S
URVIVED
.*
V
ARIABLE
N
ORMAL
S
UBJECTS
S
UBJECTS
WITH
C
ARDIOVASCULAR
D
ISEASE
TOTAL
(
N
=2534)
SURVIVED
(
N
=2246)
DIED
(
N
=288) P
VALUE
TOTAL
(
N
=3679)
SURVIVED
(
N
=2711)
DIED
(
N
=968) P
VALUE
Age (yr) 55±12 55±12 62±10 <0.001 61±10 60±10 65±9 <0.001
Height (in.) 69.4±3.4 69.4±3.1 69.7±4.9 0.08 69.2±3.7 69.2±3.5 69.3±4.2 0.34
Weight (lb) 193.7±37.5 194.1±37.1 191.0±40.1 0.19 188.8±36.1 190.7±36.3 183.7±34.4 <0.001
Body-mass index 28.3±5.1 28.4±5.1 27.5±5.0 0.005 27.8±5.0 28.1±5.0 26.9±4.7 <0.001
Resting values
Heart rate (beats/min) 78±16 78±16 83±16 <0.001 78±26 77±29 79±16 0.24
Blood pressure (mm Hg)
Diastolic 84±12 84±12 83±13 0.16 82±18 82±19 80±12 <0.001
Systolic 132±20 133±20 131±21 0.13 134±23 135±23 132±24 <0.001
Maximal values
Heart rate (beats/min) 145±24 145±23 140±25 <0.001 132±29 133±28 127±32 <0.001
Blood pressure (mm Hg)
Diastolic 86±16 86±15 85±16 0.37 86±23 86±20 85±30 0.53
Systolic 184±28 184±27 178±32 <0.001 174±31 176±31 168±32 <0.001
Exercise capacity (MET) 9.5±3.8 9.7±3.7 8.4±3.5 <0.001 7.2±3.3 7.4±3.3 6.5±2.8 <0.001
796
·
N Engl J Med, Vol. 346, No. 11
·
March 14, 2002
·
www.nejm.org
The New England Journal of Medicine
ed in Table 3. After adjustment for age, the best pre-
dictor of an increased risk of death among normal sub-
jects was peak exercise capacity, followed by pack-years
of smoking. Among subjects with cardiovascular dis-
ease, the best predictor of an increased risk of death
from any cause was peak exercise capacity, followed by
a history of congestive heart failure, history of myo-
cardial infarction, pack-years of smoking, left ventric-
ular hypertrophy on electrocardiography while at rest,
pulmonary disease, and exercise-induced ST-segment
depression. According to the model for the total
group, every 1-MET increase in exercise capacity
conferred a 12 percent improvement in survival.
The age-adjusted relative risks of death for subjects
with each of the major risk factors among those achiev-
ing a peak exercise capacity of less than 5 MET and
5 to 8 MET, as compared with the fittest subjects
(those achieving a peak of more than 8 MET), are
shown in Figure 1. For subjects with any of these risk
factors, the relative risk of death from any cause in-
creased significantly as exercise capacity decreased. The
age-adjusted relative risks of death from any cause for
subjects in each quintile of fitness in each group are
shown in Figure 2. In both groups, subjects with low-
er exercise capacity had a higher risk of death. The rel-
ative risk for the subjects in the lowest quintile of ex-
ercise capacity, as compared with those in the highest
quintile, was 4.5 among the normal subjects and 4.1
among those with a history of cardiovascular or pul-
monary disease, abnormal results on exercise testing,
or both.
Absolute Exercise Capacity versus Percentage
of Age-Predicted Value
Absolute peak exercise capacity (with or without
adjustment for age) predicted survival more accurately
than the percentage of age-predicted values achieved
when entered into the proportional-hazards model. In
addition, the area under the receiver-operating-char-
acteristic curve was greater for absolute exercise ca-
pacity than for the percentage of age-predicted values
(0.67 vs. 0.62, P<0.01), indicating that the absolute
value had greater discriminatory power. For subjects
over 65 years of age, however, the areas under the
receiver-operating-characteristic curves were similar
(0.60). The survival curves for normal subjects who
achieved an exercise capacity of less than 5 MET, 5 to
8 MET, and more than 8 MET are shown in Figure
3A; the survival curves for normal subjects who
achieved an exercise capacity of less than 50 percent,
50 to 74 percent, 75 to 100 percent, and more than
100 percent of the age-predicted value are shown in
Figure 3B. The corresponding curves for the subjects
with cardiovascular disease are shown in Figures 3C
and 3D. For both the absolute exercise capacity and
the percentage of the age-predicted value, there were
significant differences in mortality rate among groups
defined according to exercise level (P<0.001), al-
though the curves were shifted downward in the group
with cardiovascular or pulmonary disease.
Effect of Beta-Blockade
There was no interaction between the use or non-
use of beta-blockade and the predictive power of the
peak exercise capacity; this was the case throughout
the typical range of values for exercise capacity (2 to
10 MET). The results were similar when subjects were
included in the beta-blockade subgroup only if they
were taking a beta-blocker and had a blunted heart-
rate response to exercise (a peak heart rate of less than
85 percent of the age-predicted value). The results
were also similar (i.e., beta-blockade had no effect)
when the survival curves were based on various cut-
*Data are from the Cox proportional-hazards model. Metabolic equiva-
lents (MET) were calculated from the peak speed and grade of the tread-
mill and were evaluated as a continuous variable. Left ventricular hypertro-
phy was defined according to electrocardiographic criteria in a resting
patient. Exercise-induced arrhythmia was defined as three or more prema-
ture ventricular contractions in succession, premature ventricular contrac-
tions that accounted for 10 percent or more of total beats during exercise,
or both. Maximal heart rate was measured at peak exercise. ST-segment de-
pression was defined as an exercise-induced change of 1.0 mm or more. CI
denotes confidence interval.
TABLE 3. AGE-ADJUSTED RISK OF DEATH, ACCORDING TO
CLINICAL AND EXERCISE-TEST VARIABLES.*
VARIABLE
HAZARD RATIO
FOR DEATH
(95% CI)
P
VALUE
Normal subjects
Peak exercise capacity (for each 1-MET in-
crement)
0.84 (0.79–0.89) <0.001
Pack-yr of smoking (for each 10-yr incre-
ment)
1.09 (1.03–1.14) <0.001
History of hypertension 0.75 (0.56–1.02) 0.07
Diabetes 1.30 (0.84–2.00) 0.24
Total cholesterol level >220 mg/dl (5.7
mmol/liter)
1.21 (0.88–1.64) 0.25
Left ventricular hypertrophy 1.22 (0.57–2.63) 0.61
Exercise-induced ventricular arrhythmia 1.14 (0.64–2.01) 0.66
Maximal heart rate (for each increment of
10 beats/min)
1.00 (0.92–1.08) 0.93
Subjects with cardiovascular disease
Peak exercise capacity (for each 1-MET in-
crement)
0.91 (0.88–0.94) <0.001
History of congestive heart failure 1.67 (1.37–2.04) <0.001
History of myocardial infarction 1.60 (1.35–1.90) <0.001
Pack-yr of smoking (for each 10-yr incre-
ment)
1.05 (1.02–1.08) 0.001
Left ventricular hypertrophy 1.50 (1.13–1.99) 0.005
Pulmonary disease 1.34 (1.06–1.68) 0.01
ST-segment depression 1.22 (1.03–1.44) 0.02
Total cholesterol level >220 mg/dl (5.7
mmol/liter)
0.88 (0.74–1.04) 0.14
Maximal heart rate (for each increment of
10 beats/min)
0.97 (0.93–1.01) 0.17
Exercise-induced ventricular arrhythmia 1.19 (0.92–1.53) 0.18
Diabetes 0.90 (0.69–1.16) 0.41
History of hypertension 1.07 (0.90–1.25) 0.47
EXERCISE CAPACITY AND MORTALITY
N Engl J Med, Vol. 346, No. 11 · March 14, 2002 · www.nejm.org · 797
off points for the percentage of age-predicted exer-
cise capacity achieved (e.g., 50 percent or 75 percent
of age-predicted values).
DISCUSSION
Our results demonstrate that exercise capacity is a
strong predictor of the risk of death in patients re-
ferred for exercise testing for clinical reasons. The
importance of exercise capacity, physical-activity sta-
tus, or both in predicting survival has been reported
in asymptomatic populations such as those of the
Framingham Study,
11
the Aerobics Center Longitu-
dinal Study,
8,9
the Lipid Research Clinics Trial,
7
and
the Harvard Alumni study.
24
Our population was
unique in that it afforded us the opportunity to assess
subjects both with and without documented cardio-
vascular disease. Whereas the above-mentioned studies
involved generally healthy populations, our data dem-
onstrate that exercise capacity is a similarly important
marker of risk in a clinically referred population and
among men both with and without existing cardiovas-
cular disease. Unlike the estimates of activity status or
the submaximal exercise tests used in many studies, the
maximal exercise testing used in the present study pro-
vided an objective measure of physical fitness.
25
In both healthy subjects and those with cardiovas-
cular disease, the peak exercise capacity achieved was
a stronger predictor of an increased risk of death than
clinical variables or established risk factors such as hy-
pertension, smoking, and diabetes, as well as other ex-
ercise-test variables, including ST-segment depression,
the peak heart rate, or the development of arrhythmias
during exercise. Our data also confirm the protective
role of a higher exercise capacity even in the presence
of other risk factors.
7-9,24, 25
In all subgroups defined
according to risk factors, the risk of death from any
cause in subjects whose exercise capacity was less than
5 MET was roughly double that of subjects whose ex-
ercise capacity was more than 8 MET (Fig. 1).
Poor physical fitness is a modifiable risk factor, and
improvements in fitness over time have been demon-
strated to improve prognosis.
2,9
Our observation that
every 1-MET increase in treadmill performance was
associated with a 12 percent improvement in survival
underscores the relatively strong prognostic value of
exercise capacity. In addition, it confirms the presence
of a graded, inverse relation between exercise capacity
and mortality from any cause.
7-11
Recent long-term
findings from the National Exercise and Heart Disease
Project
26
among patients who had had a myocardial
Figure 1. Relative Risks of Death from Any Cause among Subjects with Various Risk Factors Who Achieved an Exercise Capacity of
Less Than 5 MET or 5 to 8 MET, as Compared with Subjects Whose Exercise Capacity Was More Than 8 MET.
Numbers in parentheses are 95 percent confidence intervals for the relative risks. BMI denotes body-mass index, and COPD chronic
obstructive pulmonary disease.
0.0
2.5
History of
Hypertension
COPD Diabetes
(1.2–1.6)
(1.7–2.3)
Smoking BMI »30 Total Cholesterol
>220 mg/dl
0.5
1.0
1.5
2.0
Risk Factors
Relative Risk of Death
>8 MET (n=2743)
5–8 MET (n=1885)
<5 MET (n=1585)
(0.8–2.1)
(1.0–2.7)
(0.9–1.9)
(1.5–3.5)
(1.1–1.6)
(1.6–2.3)
(1.2–2.0)
(1.8–3.0)
(1.2–1.8)
(1.6–2.3)
798 · N Engl J Med, Vol. 346, No. 11 · March 14, 2002 · www.nejm.org
The New England Journal of Medicine
infarction demonstrated that every 1-MET increase in
exercise capacity after a training period was associat-
ed with a reduction in mortality from any cause that
ranged from 8 percent to 14 percent over the course
of 19 years of follow-up. In a study involving serial
evaluations in nearly 10,000 men, Blair et al.
9
observed
a 7.9 percent reduction in mortality for every one-min-
ute increase in treadmill time (roughly equivalent to
the 1-MET change in our study).
In combination, these findings demonstrate that
both a relatively high degree of fitness at base line and
an improvement in fitness over time yield marked re-
ductions in risk. The relative weight of exercise capac-
ity in the model for assessing risk in both normal
subjects and those with cardiovascular or pulmonary
disease in our study, along with the fact that an im-
provement in exercise capacity lowers the risk of
death,
9,26
suggests that health professionals should
incorporate into their practices strategies for promot-
ing physical activity, in addition to the routine treat-
ment of hypertension and diabetes, the encourage-
ment of smoking cessation, and the like.
Our findings in normal subjects are similar to those
of other studies
8,27,28
in that we observed the most
striking difference in mortality rates between the least-
fit quintile and the next-least-fit quintile. This obser-
vation concurs with the consensus (reflected in the
recommendations of the Centers for Disease Control
and Prevention and the American College of Sports
Medicine
2
and the report of the Surgeon General on
physical activity and health
29
) that the greatest health
benefits are achieved by increasing physical activity
among the least fit. Among subjects with cardiovascu-
lar disease, however, we observed a nearly linear reduc-
tion in risk with increasing quintiles of fitness. Since
most studies assessing the relation between fitness and
mortality have excluded subjects with cardiovascular
disease,
30
these findings require confirmation.
Few studies have similarly assessed the prognostic
value of exercise tolerance among patients specifically
referred for exercise testing for clinical reasons. Roger
et al.
31
retrospectively assessed 2913 men and women
from Olmsted County, Minnesota, and reported that
among exercise-test variables, exercise capacity had the
strongest association with overall mortality and cardi-
ac events among subjects of both sexes. More recently,
this group addressed the association between clinical
and exercise-test variables among young and elderly
subjects in Olmsted County and observed that the
peak workload achieved was the only treadmill-test
Figure 2. Age-Adjusted Relative Risks of Death from Any Cause According to Quintile of Exercise Capacity among Nor-
mal Subjects and Subjects with Cardiovascular Disease.
The subgroup of subjects with the highest exercise capacity (quintile 5) was used as the reference category. For each
quintile, the range of values for exercise capacity represented appears within each bar; 95 percent confidence intervals
for the relative risks appear above each bar.
0.0
5.0
1234
5
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
Quintiles of Exercise Capacity
Relative Risk of Death
(3.0–6.8)1.0–5.9 MET
(3.3–5.2)1.0–4.9 MET
(1.5–3.8)6.0–7.9 MET
(2.4–3.7)5.0–6.4 MET
(1.1–2.8)8.0–9.9 MET
(1.7–2.8)6.5–8.2 MET
(0.7–2.2)10.0–12.9 MET
(1.4–2.2)8.3–10.6 MET
»13.0 MET
»10.7 MET
Normal subjects
Subjects with cardiovascular disease
EXERCISE CAPACITY AND MORTALITY
N Engl J Med, Vol. 346, No. 11 · March 14, 2002 · www.nejm.org · 799
Figure 3. Survival Curves for Normal Subjects Stratified According to Peak Exercise Capacity (Panel A) and According to the Per-
centage of Age-Predicted Exercise Capacity Achieved (Panel B) and Survival Curves for Subjects with Cardiovascular Disease Strat-
ified According to Peak Exercise Capacity (Panel C) and According to the Percentage of Age-Predicted Exercise Capacity Achieved
(Panel D).
In all the analyses, the stratification according to exercise capacity discriminated among groups of subjects with significantly dif-
ferent mortality rates — that is, the survival rate was lower as exercise capacity decreased (P<0.001).
0
100
B
Normal Subjects
0 3.5 7.0 10.5 14.0
75
50
25
Years of Follow-up
>100%
75–100%
50–74%
<50%
Percentage Surviving
0
100
A
Normal Subjects
0 3.5 7.0 10.5 14.0
75
50
25
>8 MET
5–8 MET
<5 MET
Percentage Surviving
0
100
D
Subjects with Cardiovascular Disease
0 3.5 7.0 10.5 14.0
75
50
25
Years of Follow-up
>100%
75–100%
50–74%
<50%
Percentage Surviving
0
100
C
Subjects with Cardiovascular Disease
0 3.5 7.0 10.5 14.0
75
50
25
>8 MET
5–8 MET
<5 MET
Percentage Surviving
variable that was significantly associated with mortal-
ity from any cause.
32
These investigators also observed
that each 1-MET increment in the peak treadmill
workload was associated with a 14 percent reduction
in cardiac events among younger subjects (those less
than 65 years old) and an 18 percent reduction among
elderly subjects.
In recent years, questions have been raised about
which variable has superior prognostic power: exercise
capacity relative to age- and sex-predicted standards
or absolute exercise capacity.
33-35
We found that exer-
cise capacity expressed as a percentage of the age-pre-
dicted value was not superior to the absolute peak ex-
ercise capacity in terms of predicting survival. Other
studies in this area have focused only on patients with
congestive heart failure and have had conflicting find-
ings.
33-35
We were also interested in whether our results
800 · N Engl J Med, Vol. 346, No. 11 · March 14, 2002 · www.nejm.org
The New England Journal of Medicine
would be affected by beta-blockade, given that such
treatment favorably influences survival and is known
to either improve or inhibit exercise tolerance, depend-
ing on the presence or absence of symptoms during
exercise, among other factors. Previous data in this
area, although sparse, have demonstrated that beta-
blockade does not interfere with the prognostic power
of a finding of a low exercise capacity.
36,37
Approxi-
mately 19 percent of the subjects in our study under-
went exercise testing while receiving beta-blocker ther-
apy, and the overall survival rate was slightly lower
among those taking beta-blockers (18.4 percent vs.
21.0 percent among those not taking such drugs,
P=0.03). Subjects achieving an exercise capacity of
5 MET or more had a higher survival rate than those
achieving an exercise capacity of less than 5 MET, and
this remained true when subjects were stratified ac-
cording to the use or nonuse of a beta-blocker. Sim-
ilarly, beta-blockade had no effect on survival within
groups of subjects stratified according to exercise ca-
pacity within the clinically relevant range (2 to 10
MET). This issue has rarely been addressed in previ-
ous studies, although presumably a substantial propor-
tion of subjects were taking beta-blockers in the Cor-
onary Artery Surgery Study,
38
the Olmsted County
Study,
31,32
the Kuopio Ischemic Heart Disease Risk
Factor Study,
10
and other follow-up studies that quan-
tified exercise tolerance and survival.
Our findings are applicable only to men, which is
noteworthy, given that exercise-test results have been
shown to differ significantly between men and wom-
en.
39
In addition, we had information only on death
from any cause; we did not know the specific causes
of death, nor were we able to censor data at the time
of cardiovascular interventions. Finally, our exercise-
capacity data were estimated on the basis of the speed
and grade of the treadmill. Although this type of es-
timate is the most common clinical measure of exercise
tolerance, directly measured exercise capacity (peak ox-
ygen consumption) is known to be a more accurate
and reproducible measure of exercise tolerance,
40
as
well as a more robust predictor of outcomes.
34,35
The present results confirm the prognostic useful-
ness of exercise capacity in men. The prognostic power
of exercise capacity is similar among apparently healthy
persons and patients with cardiovascular conditions
who are referred for exercise testing and similar among
subjects who are taking beta-blockers and those who
are not taking beta-blockers. Expressing exercise ca-
pacity as a percentage of the age-predicted value does
not improve its prognostic power. Our findings dem-
onstrate an association between exercise capacity and
overall mortality, not necessarily a causal relation. Nev-
ertheless, given the high prognostic value of exercise
capacity relative to other markers of risk in this and
other recent studies, clinicians who are reviewing ex-
ercise-test results should encourage patients to improve
their exercise capacity. In terms of reducing mortality
from any cause, improving exercise tolerance warrants
at least as much attention as other major risk factors
from physicians who treat patients with or at high
risk for cardiovascular disease.
REFERENCES
1. Chang JA, Froelicher VF. Clinical and exercise test markers of prognosis
in patients with stable coronary artery disease. Curr Probl Cardiol 1994;
19:533-87.
2. Pate RR, Pratt M, Blair SN, et al. Physical activity and public health:
a recommendation from the Centers for Disease Control and Prevention
and the American College of Sports Medicine. JAMA 1995;273:402-7.
3. Froelicher VF, Myers J. Exercise and the heart. 4th ed. Philadelphia:
W.B. Saunders, 2000.
4. Gibbons RJ, Balady GJ, Beasley JW, et al. ACC/AHA guidelines for ex-
ercise testing: a report of the American College of Cardiology/American
Heart Association Task Force on Practice Guidelines (Committee on Ex-
ercise Testing). J Am Coll Cardiol 1997;30:260-311.
5. Froelicher V. Exercise testing in the new millennium. Prim Care 2001;
28:1-4.
6. Froelicher VF, Fearon WF, Ferguson CM, et al. Lessons learned from
studies of the standard exercise ECG test. Chest 1999;116:1442-51.
7. Ekelund LG, Haskell WL, Johnson JL, Whaley FS, Criqui MH, Shops
DS. Physical fitness as a predictor of cardiovascular mortality in asympto-
matic North American men: the Lipid Research Clinics Mortality Follow-
up Study. N Engl J Med 1988;319:1379-84.
8. Blair SN, Kohl HW III, Paffenbarger RS Jr, Clark DG, Cooper KH,
Gibbons LW. Physical fitness and all-cause mortality: a prospective study of
healthy men and women. JAMA 1989;262:2395-401.
9. Blair SN, Kohl HW III, Barlow CE, Paffenbarger RS Jr, Gibbons LW,
Macera CA. Changes in physical fitness and all-cause mortality: a prospec-
tive study of healthy and unhealthy men. JAMA 1995;273:1093-8.
10. Lakka TA, Venäläinen JM, Rauramaa R, Salonen R, Tuomilehto J, Sa-
lonen JT. Relation of leisure-time physical activity and cardiorespiratory fit-
ness to the risk of acute myocardial infarction in men. N Engl J Med 1994;
330:1549-54.
11. Kannel WB, Wilson P, Blair SN. Epidemiological assessment of the
role of physical activity and fitness in development of cardiovascular dis-
ease. Am Heart J 1985;109:876-85.
12. Froelicher VF, Myers J. Research as part of clinical practice: use of
Windows-based relational data bases. Veterans Health System Journal.
March 1998:53-7.
13. Froelicher VF, Shiu P. Exercise test interpretation system. Physicians
and Computers 1996;14:40-4.
14. Wolthuis R, Froelicher VF Jr, Fischer J, et al. New practical treadmill
protocol for clinical use. Am J Cardiol 1977;39:697-700.
15. Myers J, Buchanan N, Walsh D, et al. Comparison of the ramp versus
standard exercise protocols. J Am Coll Cardiol 1991;17:1334-42.
16. Myers J, Do D, Herbert W, Ribisl P, Froelicher VF. A nomogram to
predict exercise capacity from a specific activity questionnaire and clinical
data. Am J Cardiol 1994;73:591-6.
17. Lachterman B, Lehmann KG, Abrahamson D, Froelicher VF. “Recov-
ery only” ST-segment depression and the predictive accuracy of the exer-
cise test. Ann Intern Med 1990;112:11-6. [Erratum, Ann Intern Med
1990;113:333-4.]
18. Yang JC, Wesley RC Jr, Froelicher VF. Ventricular tachycardia during
routine treadmill testing: risk and prognosis. Arch Intern Med 1991;151:
349-53.
19. American College of Sports Medicine. Guidelines for exercise testing
and prescription. 6th ed. Baltimore: Lippincott Williams & Wilkins, 2000.
20. Morrow K, Morris CK, Froelicher VF, et al. Prediction of cardiovas-
cular death in men undergoing noninvasive evaluation for coronary artery
disease. Ann Intern Med 1993;118:689-95.
21. Reid MC, Lachs MS, Feinstein AR. Use of methodological standards
in diagnostic test research: getting better but still not good. JAMA 1995;
274:645-51.
22. Shue P, Froelicher VF. EXTRA: an expert system for exercise test re-
porting. J Non-Invasive Testing 1998;II-4:21-7.
23. Morris CK, Myers J, Froelicher VF, Kawaguchi T, Ueshima K, Hideg
A. Nomogram based on metabolic equivalents and age for assessing aero-
bic exercise capacity in men. J Am Coll Cardiol 1993;22:175-82.
EXERCISE CAPACITY AND MORTALITY
N Engl J Med, Vol. 346, No. 11 · March 14, 2002 · www.nejm.org · 801
24. Paffenbarger RS Jr, Hyde RT, Wing AL, Hsieh CC. Physical activity,
all-cause mortality, and longevity of college alumni. N Engl J Med 1986;
314:605-13.
25. Haskell WL, Leon AS, Caspersen CJ, et al. Cardiovascular benefits and
assessment of physical activity and physical fitness in adults. Med Sci Sports
Exerc 1992;24:Suppl:S201-S220.
26. Dorn J, Naughton J, Imamura D, Trevisan MA. Results of a multi-
center randomized clinical trial of exercise and long-term survival in myo-
cardial infarction patients. Circulation 1999;100:1764-9.
27. Blair SN, Jackson AS. Physical fitness and activity as separate heart dis-
ease risk factors: a meta analysis. Med Sci Sports Exerc 2001;33:762-4.
28. Williams PT. Physical fitness and activity as separate heart disease risk
factors: a meta-analysis. Med Sci Sports Exerc 2001;22:754-61.
29. Physical activity and health: a report of the Surgeon General. Atlanta:
National Center for Chronic Disease Prevention and Health Promotion,
1996.
30. Blair SN. Physical activity, fitness and coronary heart disease. In: Bou-
chard C, Shepard RJ, Stephens T, eds. Physical activity, fitness and health.
Champaign, Ill.: Human Kinetics, 1994:579-90.
31. Roger VL, Jacobsen SJ, Pellikka PA, Miller TD, Bailey KR , Gersh BJ.
Prognostic value of treadmill exercise testing: a population-based study in
Olmsted County, Minnesota. Circulation 1998;98:2836-41.
32. Goraya TY, Jacobsen SJ, Pellikka PA, et al. Prognostic value of tread-
mill exercise testing in elderly persons. Ann Intern Med 2000;132:862-
70.
33. Aaronson KD, Mancini DM. Is percentage of predicted maximal exer-
cise oxygen consumption a better predictor of survival than peak exercise
oxygen consumption for patients with severe heart failure? J Heart Lung
Transplant 1995;14:981-9.
34. Stelken AM, Younis LT, Jennison SH, et al. Prognostic value of car-
diopulmonary exercise testing using percent achieved of predicted peak ox-
ygen uptake for patients with ischemic and dilated cardiomyopathy. J Am
Coll Cardiol 1996;27:345-52.
35. Myers J, Gullestad L, Vagelos R, et al. Clinical, hemodynamic, and
cardiopulmonary exercise test determinants of survival in patients referred
for evaluation of heart failure. Ann Intern Med 1998;129:286-93.
36. Ronnevik RK, von der Lippe G. Prognostic importance of predis-
charge exercise capacity for long-term mortality and non-fatal myocardial
infarction in patients admitted for suspected acute myocardial infarction
and treated with metoprolol. Eur Heart J 1992;13:1468-72.
37. Murray DP, Tan LB, Salih M, Weissberg P, Murray RG, Littler WA.
Does beta adrenergic blockade influence the prognostic implications of
post-myocardial infarction exercise testing? Br Heart J 1988;60:474-9.
38. Weiner DA, Ryan TJ, Parsons L, et al. Long-term prognostic value of
exercise testing in men and women from the Coronary Artery Surgery
Study (CASS) registry. Am J Cardiol 1995;75:865-70.
39. Shaw LJ, Hachamovitch R , Redberg RF. Current evidence on diagnos-
tic testing in women with suspected coronary artery disease: choosing the
appropriate test. Cardiol Rev 2000;8:65-74.
40. Myers J. Essentials of cardiopulmonary exercise testing. Champaign,
Ill.: Human Kinetics, 1996.
Copyright © 2002 Massachusetts Medical Society.
POSTING PRESENTATIONS AT MEDICAL MEETINGS ON THE INTERNET
Posting an audio recording of an oral presentation at a medical meeting on the
Internet, with selected slides from the presentation, will not be considered prior
publication. This will allow students and physicians who are unable to attend the
meeting to hear the presentation and view the slides. If there are any questions
about this policy, authors should feel free to call the Journal’s Editorial Offices.
