Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

CURRICULUM IN CARDIOLOGY

The prognostic value of exercise capacity: A


review of the literature
Charles K. Morris, MD, Kenji Ueshima, MD, Takeo Kawaguchi, MD,
Alisa Hideg, and Victor F. Froelicher, MD. Long Beach, Calif.

Defining noninvasive parameters that are predictive basal oxygen requirement to maintain life in the
of cardiovascular mortality has long been a goal of resting state is approximately 3.5 ml O&g/min, this
researchers and clinicians alike. Since Ellestad and has been designated as 1 MET (metabolic equiva-
Wan’s pioneering work in 1975l up to the present, lent), and multiples thereof have been used to de-
there has been a plethora of prospective and retro- scribe levels of VO2 max (i.e., 35 ml OJkg/min = 10
spective studies endeavoring to sort out the “wheat METs).~ As most treadmill protocols proceed in spe-
from the chaff” in exercise testing parameters. Exer- cific timed steps (mph/% grade), exercise time re-
cise capacity, as estimated by exercise duration or lates to oxygen consumption (METS), although the
work load during an exercise test, has been thought accuracy of this relationship has been questioned.5
by clinicians to be a significant predictor of cardio- The advantage of reporting results in terms of METS
vascular mortality. This review discussesthe signif- rather than in minutes is that different protocols can
icant research in this field and attempts to define the thereby be more meaningfully compared and exercise
relevance of exercise capacity. test performance can be related to everyday activi-
Exercise work load has been well correlated with ties.
maximal ventilatory oxygen uptake (VOa max) for The questions that we shall discussthat have been
the various treadmill and bicycle protocols within a addressed by the various studies include: (1) Is exer-
known margin of error. But over the years, debate has cise capacity an independent predictor of cardiovas-
continued as to the validity of estimation versus cular mortality? (2) Is exercise capacity an indepen-
measurement using expired gas analysis of VOa max. dent predictor of all causes of death? (3) Are there
Using a single exercise test as a measure of a patient’s particular MET levels that can predict differences in
exercise capacity has also been controversial. Most survival?
authorities will readily admit the shortcomings of the
CARDIOVASCULAR MORTALITY
exercise test, yet it has nonetheless become recog-
nized not only as the most convenient form of assess- We excluded from our analysis studies of exercise
ment, but as the most quantitative and reproducible capacity involving patients who had suffered recent
as well. While it may not indicate overall physical myocardial infarctions. The reasons for this were the
fitness, it does correlate with maximal aerobic capac- frequency of submaximal treadmill tests in such
ity, which relates well to maximal cardiac output. studies, which may not give an accurate assessment
The aerobic cost or VOz max (in ml Op/kg/min) of of true exercise capacity, and the high mortality rate
treadmill exercise can be calculated by means of the already inherent in this group. Certainly this sub-
following formula2: VOZ max = (mph x 26.8) x group has already been addressed in the literature. In
(0.1 + [ % grade x 1.81) t 3.5. This assumesthat the a me&analysis of the prognostic value of exercise
individual is not being carried along by the handrails testing in patients after a recent myocardial infarc-
or relying on anaerobic metabolism.3 Since a person’s tion, Froelicher et al6 found that 12 of 14 studies that
considered exercise capacity showed an increased
risk for those patients with a poor exercise capacity.
From the Cardiology Section, Long Beach Veterans Administration Med- Also excluded from review were four studies of
ical Center.
Received for publication March 20, 1991; accepted May 9, 1991.
asymptomatic populations that used exercise testing
Reprint requests: Charles K. Morris, MD, Cardiology Section (lllc), Long as a measure of exercise capacity and examined its
Beach VA Medical Center, 5901 East Seventh St., Long Beach CA 90822. relationship to cardiovascular mortality. These stud-
4/l/32462 ies (i.e., Lipid Research Clinics [LRC], Multiple Risk

1423
November 199 1
1424 Morris et al. Ametkan Heart Journal

Factor Intervention Trial [MRFIT], LA County fined by Bruce’* in 1971 as “the difference between
Workers, and Aerobic Research in Dallas7-‘O) all observed VOz max and that expected for a healthy
found a positive correlation between poor exercise person of similar age, sex and habitual activity
capacity and cardiovascular mortality. status.”
Of the studies we reviewed, 12 addressed the issue Hammermeister et al.‘” in 1979 analyzed retro-
of whether exercise capacity was a predictor of car- spectively a separate group from the Seattle Heart
diovascular death. Ellestad and Wan’ conducted a Watch Study made up of 2603 patients who were
retrospective study of 2700 patients referred for a separated into two nonrandomized groups, coronary
variety of reasons to their treadmill laboratory who artery bypass surgery versus medical treatment. Fol-
were then followed from 6 months to 9 years (no mean low-up averaged 3.6 years. Only 46% of these pa-
follow-up given) for subsequent cardiac morbidity or tients underwent treadmill testing. Multivariate
mortality. The main focus of this study was on the analysis of the exercise test variables alone found a
predictive capacity of ST segment changes. Their low double product and a low treadmill time to be
main finding was that an early onset of ischemic predictive of increased mortality. But when the
electrocardiographic changes predicted a higher mor- exercise variables were considered multivariately
tality rate and number of cardiac events. They indi- with both the clinical and angiographic variables, no
rectly assessed exercise capacity, finding that “chro- exercise variables were found to be independently
notropic incompetence” (the pulse rate is below the predictive of mortality. It should be noted, however,
95% confidence interval for a given work load, age, that they excluded treadmill variables from their fi-
and sex) forecasted a significant increase in myocar- nal multivariate analysis because of the low percent-
dial infarctions, the progression of angina, and deaths. age of patients having undergone treadmill testing,
McKirnan et al.‘l have reported subsequently that and this certainly weakens any conclusions made
maximal heart rate is the main determinant of from this analysis.
treadmill work load achieved; thus chronotropic in- McNeer et all6 published a retrospective study in
competence may be inferred to reflect a reduction in 1979 of 1472 patients who had significant coronary
exercise capacity. This increased risk was equivalent disease by angiography, an exercise test, and received
to that found in patients who had positive ST either medical or surgical therapy (nonrandomized).
segments. It should be noted that while this study Multivariate analysis of the medically treated pa-
used a chi square analysis, no multivariate or survival tients showed an abnormal ST segment response and
analysis techniques were used. a low exercise stage to be significant independent
Bruce et a1.l’ in 1977 published a prospective study predictors of mortality after 4 years of follow-up. Of
of 1852 men from the Seattle Heart Watch Study who great interest was their finding that patients who had
had a clinical history of coronary artery disease. a normal ST segment response and achieved stage IV
These men were then followed during medical treat- of a Bruce protocol (13 METS) or higher had a near
ment for an average of 3 years. They found conclu- 100% 4-year survival. Patients with a positive ST
sively that a poor exercise capacity (as measured by segment response who only achieved stage I of a
a low duration of exercise) predicted a higher inci- Bruce protocol (5 METS) or less had a high preva-
dence of sudden cardiac death. This was valid using lence of triple-vessel/left main disease (73% and
both univariate and multivariate analysis of clinical 27 % , respectively) and had a very poor prognosis (4-
and exercise data. In a subsequent study published in year survival rate less than 60 % ).
1983,13 they prospectively followed 3611 men and 547 Platia et al.” retrospectively studied 78 patients
women, free of clinically evident coronary artery dis- who received coronary angiograms between 1960 and
ease at enrollment, for a total of 10 years. In a sub- 1967 and had a greater than 70% stenosis of at least
category of asymptomatic men without atypical chest one coronary artery and were “suitable for surgery”
pain or hypertension, they discovered that any single by 1978 standards but received medical therapy at
clinical risk factor (age >55, hypertension, any his- the time. Sixty-eight of these patients performed a
tory of tobacco use, male gender) when combined Master double two-step test that was halted for
with two or more exercise variables (angina, heart symptoms or the attainment of the predicted number
rate <90% predicted, double product <80% pre- of steps. Follow-up was a mean of 9.9 years. Multi-
dicted, functional aerobic impairment >20%, func- variate analysis of clinical, exercise, and angiographic
tional age >55 years) predicted a thirty-threefold in- variables did not find exercise test time to be predic-
crease in cardiac events (worsened angina, myocar- tive of survival. It should be noted, however, that this
dial infarction, coronary artery bypass surgery, was a small study using the step test, which did not
death). Functional aerobic impairment is a term de- permit estimation of exercise capacity.
Volume 122
Number 5 Prognostic value of exercise capacity 1425

Podrid et all8 published a prospective study in tients with no prior coronary artery bypass surgery or
1981 on 14‘2 male patients with coronary artery dis- recent myocardial infarction who underwent coro-
ease by history or by cardiac catheterization who had nary angiography, left ventriculography, and exercise
>2 mm ST segment depression on at least two suc- testing within 1 month of each other. Follow-up was
cessive Bruce protocol treadmill tests. These patients a minimum of 3 years (average follow-up not stated),
received medical management and were followed an alJ it was 99% complete. The 1474 patients who un-
average of 59 months. The patients were divided into derwent coronary artery bypass surgery during fol-
three categories during follow-up: group 1 were those low-up were withdrawn at that time (i.e., censored).
who exercised 1 to 6 minutes (average 4.2 minutes, 6 With multivariate analysis of 18 clinical and seven
METS); group 2 exercised 6 to 9 minutes (average 7.5 exercise parameters, the four that were found to be
minutes, 8 METS); and group 3 exercised >9 minutes independently predictive of mortality included: left
(average 10.2 minutes, 11 METS). There was a ventricular enlargement, congestive heart failure
significant decrease in survival in group 1 compared score, ST segment response, and final exercise stage
with both groups 2 and 3. No difference in survival achieved.
existed between groups 2 and 3 themselves. These investigators 21 also divided their patients
In a prospective study of 147 men and women at into some interesting subgroups for multivariate
the National Institute of Health published in 1982, analysis of clinical, exercise, and cardiac catheteriza-
Kent et al.lg looked at the predictive capacity of ex- tion variables. In those with preserved left ventricu-
ercise testing for short-term mortality (1 year) in pa- lar function, the independent predictors of mortality
tients with known coronary artery disease by angiog- were: congestive heart failure score and peak exercise
raphy (>50% stenosis of a major coronary artery). heart rate in patients with single-vessel disease, con-
This group of patients had only mild or nonexistent gestive heart failure score in patients with double-
angina, were younger than 65, did not have signifi- vessel disease, and exercise capacity in patients with
cant left main stenosis, and had a left ventricular triple-vessel disease. In the latter group-those with
ejection fraction >20 % . They all received medical preserved left ventricular function and triple-vessel
therapy and were followed an average of 2 years disease-survival was 100 % for those who exercised
(range 6 to 67 months). Exercise capacity was not to stage V (>15 METS) or greater in the Bruce pro-
found to predict short-term mortality, except in the tocol but only 53 % for those who could only achieve
subgroup of patients with triple-vessel disease. In stage l/2 (2 METS). In patients with poor left
this subgroup, an exercise capacity of <lOO W (2 ventricular function, exercise capacity was not a sig-
METS) on an upright bicycle ergometer doubled the nificant predictor of survival. In their final multi-
mortality rate of those able to go more than 100 W. variate survival model of all patients together,21
This difference became statistically significant at 1 seven factors proved to be significant: left ventricu-
year. lar contraction pattern, number of diseased coronary
Sawada et aL2’ prospectively followed 170 pa- vessels, male gender, diuretic use, digitalis use, low
tients with normal resting and exercise echocardio- peak heart rate, and final exercise stage achieved.
grams. These patients were followed medically for Many of these factors, it should be noted, are either
an average of 28.4 months until they died or had a directly or indirectly related to left ventricular func-
cardiac event (percutaneous transluminal coronary tion.
angioplasty, coronary artery bypass surgery, or In a paper published in 1986,22 the same investiga-
myocardial infarction). Follow-up was 87% com- tors looked at an additional subclassification of 5303
plete. Their results, published in 1990, were that patients based on exercise capacity with respect to
cardiac events were far more frequent in those pa- medical versus surgical survival. It should be noted
tients who achieved <6 METS (8.9% versus 1.1% ) that patients were not randomized into the treatment
or <85% maximal predicted heart rate (10.9 % categories and that patients in these categories were
versus 1.0%). In addition, they concluded that a by no means equivalent (i.e., the medically treated
patient with a normal exercise echocardiogram and patients had significantly greater prevalence of se-
a good exercise capacity has an excellent short-term vere congestive heart failure, prior acute myocardial,
prognosis, even if he exhibits exercise-induced digitalis therapy, and worse left ventricular function
ST segment depression. by cardiac catheterization; the surgically treated pa-
A great deal of work has been done in this area by tients had a higher prevalence of an ischemic ST re-
Weiner et al. with groups of patients from the Coro- sponse to exercise, class 3 or 4 angina, and a lower
nary Artery Surgery Study (CASS) registry. In 1984 exercise capacity). After 7 years of follow-up, pa-
theyzl published a retrospective study of 4083 pa- tients who had achieved < stage I (<5 METS) had a
November 1991
1426 Morris et al. American Heart Journal

Table 1. Cardiovascular mortality


C” 0;
POP Mean v;, Study P; Pr 5% r,
Author Year size age Women Test type MI CHF Dig BB

Ellestad’ 1975 2761 50 18 TM Retro - --

Bruce12 1977 1852 53 0 TM Prosp - - - -


McNeer”j 1978 1472 - - TM Retro - - - -
Hammermeister15 1979 - 14 TM Retro - - - 100

Podrid’* 1981 142 57 0 TM Prosp 66 0 0 44 88

Kentlg 1982 147 49 10 Bike Prosp - - - 82

Bruce13 1983 4158 46 13 TM Prosp 0 0 0

Weinerzl 1984 - 23 TM Retro - - - - 100

Weinerz2 1986 5303 - 22 TM Retro 47 3 10 40 100

Dagenais25 1988 298 52 9 TM Retro 32 - - - 265

Weiner23 1987 1249 - 8 TM 60 5 13 44 100

Bogaty2” 1989 241 50 25 TM Retro 35 0 0 - -


Mark24 1989 2842 49 30 TM Prosp 29 4 -- - 100

SawadaeO 1990 170 53 48 TM Prosp 0 0 - - 0

Pr, prior; MI, myocardial infarction; CHF, congestive heart failure; Dig, digitalis; BB, /3-blockers; CAD, coronary artery disease; Cath, catheterization; FU,
follow-up; CABS, coronary artery bypass surgery; CV, cardiovascular; ExCap, exercise capacity; TM, treadmill; Retro, retrospective; Prosp, prospective; Chi
so; chi square analysis; KM, Kaplan-Meier; Ext, (Fisher) exact test; 3V, three-vessel; combo, combination; U, univariate; M, multivariate; Exp, expired.

significantly better survival with surgical therapy that was highly predictive of survival, even when
(81% versus 78%, p = 0.04), especially if they also considered with coronary anatomy and left ventric-
had ST segment depression with exercise (82 % ver- ular function. In the composition of this treadmill
sus 72 % , p = 0.001). Those achieving > stage II (>8 score, after consideration of ST segment response,
METS) had equal survival with either therapy. (exercise capacity as determined by treadmill time
In 1987 Weiner et al.23 published a study that ex- and treadmill angina index were the only other inde-
amined 1249 patients with triple-vessel disease who pendent predictors of mortality. In the final model,
again received either medical or surgical therapy. In- treadmill exercise time had a relative weight of 1
dependent predictors of survival in this group were: compared with 4 for each of the other variables.
left ventricular score, final exercise stage achieved, Dagenais et al. 25 retrospectively examined 298 pa-
and treatment received (surgical versus medical). At tients with exercise tests exhibiting at least 2 mm of
Duke University, Mark et al.24 in 1987 published a ST segment depression who were treated medically.
significant prospective study of 2642 patients with They contrasted the survival in those with and with-
suspected coronary artery disease by history who re- out exercise-induced angina. Of interest, not only did
ceived both cardiac catheterization and a treadmill they find that the silent ischemia group had a better
test within 6 weeks of each other. The main result of survival than the group with angina Cp = 0.05), but
this study was the development of a treadmill score they also concluded that survival in both groups was
Volume 122
Number 5 Prognostic value of exercise capacity 1427

% CAD Mean Range % cv Low ExCap High ExCap


Cath by FU FU % % FU % FU FU Study increases decreases
REQD Cath (Yr) (Yd FU CABS death death analysis mortality mortality

No - - 0.5-9.0 100 5.0 13.2 Life Chronotropic


table, incompetence
chi sq increases
mortality
No - 2.8 1.7-3.8 100 - 10.5 7.7 UN Yes
Yes 59 4.0 - 95 21.0 - - UN Yes
Yes 88 3.6 2.5-8.2 100 72 9.9 8.5 U,M, No, but only
cox 46% had TM
No 12 4.9 1.6-8.3 100 6.3 7.7 7.0 cox, Yes
life table
Yes - 2.4 - 81 2.7 5.4 5.4 KM No, x3V
disease
No n 6.1 - 93 - 2.6 1.0 UN Yes, In combo Chronotropic
with clinical incompetence
risk factor increases
mortality
Yes - - 3.0-5.0 99 36.0 5.2 4.3 U,M Yes, Exp
cox in 3V
disease
Yes - - 4.0-7.0 99 30.9 10.6 8.6 UM, Survival
cox, improved
chi sq by surgery
No 769 - 76.0 100 0 18 16 KM, Yes Yes
Cox, M
Yes - - 5.0-7.0 99 62.4 - - U&f, Poor Ex Cap
cox do better
with surgery
Yes 100 - 1.3-8.0 100 21.1 21.6 18.2 Yes Yes
Yes 61 5 -10.0 98 24.0 9.5 - Cox, KM, Yes Yes
Duke score
No 0 2.4 - 87 2.7 0 0 Fisher Yes
Ext,
t test

strongly correlated with exercise duration. (In the si- cardiac mortality only. Of the 52 deaths in the study
lent ischemia group, survivors exercised 472 k 129 by Bogaty et al.,26only six were from medical causes
secondswhile nonsurvivors exercised 360 + 159 sec- other than cardiac. Bruce et al. did not comment on
onds [p < 0.011.) Survival was nearly 100% in those the extremely high percentage of noncardiac deaths
who reached 10 METS in either group. With multi- in both of their studies, 41% (80 of 195) of deaths in
variate analysis of the 194 patients who had also un- 197712and 63% (68 of 107) in 1983.13In their 1982
dergone coronary angiography, only the presence of study, Dagenais et al.27reported only 5 of 54 deaths
multivessel disease and a low exercise duration were (9%) being noncardiac, and in 1990 they25 again re-
predictive of decreased survival. A subsequent paper ported only 3 of 32 deaths (9 % ) as being noncardiac.
published from the same institution, analyzing a Harris et a1.28reported only 14 of 259 deaths as non-
similar group of subjects, stated that exercise capac- cardiac (5% ) in their study published in 1979 of 1214
ity is predictive of survival but not of morbid events patients with coronary artery diseaseby angiography
including myocardial infarction or unstable angina.26 who were treated medically and followed for at least
A summary of these studies is presented in Table I. 1year. Kent et all9 had-only eight deaths over 2 years
of follow-up, all of which were cardiac. Weiner et al.
ALL-TYPE MORTALITY
did not report any details of their noncardiac deaths
The majority of studies previously cited do not in any of their studies, but in the one published in
discuss the prediction of noncardiovascular mortal- 198421they did state that 18% (38 of 212) of the
ity, and the majority of their survival data is based on deaths were in this category. So it would seem that
November 199 1
1428 Morris et al. American Heart Jotwnal

Table II. Survival in patients with a markedly positive (2 2 in medically treated patients with coronary artery
mm ST depression)exercisetest disease to be predictive of increased sudden cardiac
-
5% 6 Year with Without exercise angina death by multivariate analysis. When Dagenais et
METS exercise angina (silent ischemia) a1.25 looked at their group of patients with “strongly
positive exercise ECGs,” they found exercise capac-
213 loo”c 97 “;a
ity to be the only significant variable predicting sur-
510 94% 87 “t
r7 80 ‘7 64 “;a vival. The nonsurvivors averaged an exercise dura-
15 44% 60 % tion of 360 t 159 seconds (6 METS) for those with-
out exercise-induced angina and 247 rf. 98 seconds (5
METS) for those with such angina. Their survivors
had an exercise capacity averaging 472 +- 129 sec-
onds (8 METS) for those without angina and
not only was there little interest in these noncardiac 344 ? 106 seconds (6 METS) for those with angina.
deaths by these particular investigators, but they In their group of medically treated patients with cor-
have also reported wide ranges of incidence of these onary artery disease, McNeer et a1.i6 determined that
events. Yet if the study groups are analyzed, it can be an exercise capacity of less than or equal to stage I of
seen that there was a range of 0 % to 10 % of noncar- Bruce (less than 5 METS) predicted not only the
disc deaths in these patients with known coronary presence of significant coronary artery disease, but
disease. Cohorts such as that of the Seattle Heart also decreased survival. Exercise capacity did not
Watch study, published by Bruce et al. in 1983,13 in- correlate with the absence of coronary artery disease
volved asymptomatic patients without known coro- though; 47 % of their patients who achieved stage IV
nary artery disease and consequently had a very high Bruce (>13 METS) had significant coronary artery
incidence of noncardiac death (i.e., 63%). disease. The most significant differences in survival
It is important to know whether a low exercise ca- were noted in their patients with coronary artery
pacity is predictive of all types of deaths, rather than disease; those exercising to greater than stage IV
just those of cardiac origin. Without this information, Bruce (13 METS) had a 24-month survival of 98%)
it is not possible to decipher whether other causes of while those exercising less than stage I Bruce (<5
death have a relationship to exercise capacity (i.e., METS) had a survival rate of only 59 % over the same
could comorbid conditions such as chronic obstruc- period.
tive pulmonary disease, cancer, malnutrition, etc., be Podrid et al-l8 divided their patients with known
confounding the data so that in reality exercise coronary disease into three groups based on exercise
capacity is not predicting coronary artery disease time and found the group that exercised an average
mortality itself?). If one is trying to predict cardiac of only 4.2 minutes (less than 6 METS) had a signif-
deaths, noncardiac deaths must be censored (i.e., icantly lower survival rate after 6.7 years than the
follow-up terminated) and not be considered as an other groups with a better exercise capacity. Of note,
actual end point. It must be understood that certain they also found no significant difference in mortality
conditions associated with noncardiac deaths could between the group with an average exercise time of
also be limiting exercise capacity. A worse case sce- 7.5 minutes (8 METS) and that with one of 10.2 min-
nario, for instance, would be if a poor exercise capac- utes (11 METS).
ity is only predictive of death in cardiac patients with Sawada et a1.,20 in their study of patient with nor-
these noncardiac-related conditions. Such questions mal exercise echocardiograms, found an exercise ca-
must be resolved for the optimal predictive use of pacity of less than 6 METS to be predictive of
exercise testing. increased cardiovascular morbidity, but they had no
deaths during their mean follow-up of 2.4 months.
MET LEVELS PREDICTIVE OF MORTALITY Weiner et a1.21 reported that in patients with pre-
Now that an overwhelming number of excellent served left ventricular function and triple-vessel dis-
studies have determined that exercise capacity is an ease, of those exercising to stage ‘/z Bruce (2 METS)
independent risk factor for cardiac death, it must be only 53 % survived 4 years, compared with 100 % for
ascertained which MET levels can be used as cutoffs those able to exercise to stage V Bruce or greater (>15
to classify patients at increased risk. Many studies METS). In another group of patients with coronary
have pronounced a low end of the spectrum, a MET artery disease followed for 7 years, these inves-
level below which survival is decreased. Bruce et al. tigators22 found that an exercise capacity of < stage
determined an exercise capacity of less than 3 METS I Bruce (<5 METS) predicted a group with a higher
Volume 122
Number 5 Prognosticvalue of exercisecapacity 1429

Table Ill. Studies relating survival to MET levels


Author Year publ. No. subjects Subject population RX Results

Bruce12 1977 1852 Males with h/o CAD Med <5 METS predicts
increase in sudden
death
McNeer“j 1978 1472 Males and females, Med <5 METS predicts
59 % with and 41% significant CAD and
without CAD by angio Surg decreased survival
> 13 METS predicts
increased survival
>13 METS does not
predict absence of CAD
Podridls 1981 142 Males with h/o CAD Med <6 METS predicts
and positive TMET Surg decreased survival
No difference in survival
in those >8 METS
WeineG 1984 4083 Males and females Med <2 METS had 53%
with CAD by angio Surg 4-year mortality
(CASS criteria) >15 METS had 100%
4-year survival
Weiner22 1986 5303 Males and females Med <5 METS predicts
with CAD by angio Surg increased mortality and
(CASS criteria) benefit from CABS
210 METS predicts
equal survival with
either Med Rx or
CABS
Dagenais25 1988 298 Males and females Med Survivors with silent
with positive TMET ischemia averaged 8
METS
Nonsurvivors with silent
ischemia averaged 6
METS
Bogaty26 1989 241 Males and females Med <5 METS had 45%
with positive TMET S-year survival rate
>lO METS had 93%
S-year survival rate
Ma&24 1987 2842 Males and females Med Risk determined by
with CAD by history treadmill score
composed of ST
segment response,
exercise time, and
exercise angina score
Sawada20 1990 173 Males and females Med Normal Ex Echo and >8
METS predicts with
normal Ex Echo Surg
good prognosis
CABS, Coronary artery bypass surgery; CAD, coronary artery disease; CASS, Coronary Artery Surgery Study; h/o, history of; Med, medical treatment; 1
MET, metabolic equivalent = 3.5 ml Oz/kg/min; Rx, recommended treatment; angio, angiography; Surg, surgical treatment; ‘IMEl’, treadmill exercisetest,;
Ex, exercise; Echo, echocardiogram.

mortality and one that benefitted from coronary ar- prognosis. A good approach would be to utilize
tery bypass surgery. receiver operating characteristic (ROC) curves to
Though exercise capacity is inversely related to age determine for different age groups the most efficient
and activity status, only Bruce et al. have adjusted MET cutpoint levels for identifying high risk.
their data for age and activity level (creating the A summary of the significant MET levels relating
Functional Aerobic Impairment [FAI] classification). to mortality is presented in Tables II and III. It is
Doing so may give physicians a better “handle” on important to note that age, activity level, or compli-
November 199 1
1430 Morris et al. Americiln Heart Journal

eating illnesses have not been considered in these serves equal emphasis with that of ST segment
studies. Age could be considered in several ways: (1) interpretation in the evaluation of exercise test
norms by decade, considering the lowest tenth per- responses.
centile as abnormal; (2) the derivation of an aerobic
impairment scale based on age; and (3) the inclusion REFERENCES
of age in a multivariate equation. For instance, it is 1. Ellestad MH, Wan MK. Predictive implications of stress test-
difficult to believe that a 5 MET threshold conveys ing. Circulation 1975;51:363-9.
the same risk for death in a 40-year-old (11 METS 2. American College of Sports Medicine. Guidelines for exercise
testing and prescription. Philadelphia: Lea & Febiger, 1986.
mean for this decade) as in a 70-year-old (7 METS 3. Froelicher VF. Exercise and the heart: Clinical concepts. Chi-
mean for this decade) or in a patient with debilitat- cago: Yearbook Medical Publishers, Inc, 1987.
ing chronic obstructive pulmonary disease. Optimal 4. Jette M, Sidney K, Blumchen G. Metabolic equivalents
(METS) in exercise testing, exercise prescription, and evalu-
use of exercise capacity as a risk predictor requires ation of functional capacity. Clin Cardiol 1990;13:555-65.
that we understand confounders so that appropriate 5. Froelicher VF, Thompson AJ, Noguera I, et al. Prediction of
MET thresholds can be chosen for appropriate clin- maximal oxygen consumption: comparison of the Bruce and
Balke treadmill protocols. Chest 1975;68:331-6.
ical subgroups. 6. Froelicher VF, Perdue S, Pewen W, Risch M. Application of
meta-analysis using an electronic spread sheet to exercise
SUMMARY testing in patients after myocardial infarction. Am J Med
1987;83:1045-54.
While there is still much debate in the literature 7. Ekelund LG, Haskell WL, Johnson JL, et al. Physical fitness
regarding the specific MET levels at which there are as a predictor of cardiovascular mortality in asymptomatic
differences in survival, the following points have be- North American men. N Engl J Med 1988;319:1379-84.
8. Leon AS, Connett J, Jacobs DR, Rauramaa R. Leisure-time
come clear with the growing body of reports in the physical activity levels and risk of coronary heart disease and
literature. Exercise capacity seems to be an indepen- death. JAMA 1987;258:2388-95.
dent predictor of mortality, and when it is combined 9. Peters RK, Cady LD, Bischoff DP, et al. Physical fitness and
subsequent myocardial infarct in healthy workers. JAMA
with other clinical, exercise, or angiographic data, it 1983;2-49:3052-6.
becomes very powerful in this regard. This relates to 10. Blair SN. Kohl HW. Paffenbareer RS. Clark DG. et al. Phvs-
both overall mortality and to that from cardiovascu- ical fitness and all ‘cause mortality: a prospective study-of
healthy men and women. JAMA 1989;262:2395-401.
lar disease. There is still a need for the establishment 11. McKirnan MD, Sullivan M, Jensen D, Froelicher VF. Tread-
of mortality data related to MET levels adjusted for mill performance and cardiac function in selected patients
age and activity status. with coronary heart disease. J Am Coil Cardiol1984;3:253-61.
12. Bruce RA, DeRouen T, Peterson DR, et al. Noninvasive pre-
A low exercise capacity of less than 6 METS indi- dictors of sudden cardiac death in men with coronary heart
cates a higher mortality group, probably regardless of disease. Am J Cardiol 1977;39:833-40.
the underlying extent of coronary disease or left ven- 13. Bruce RA, Hossack KF, DeRouen TA, Hofer V. Enhanced risk
assessment for primary coronary heart disease events by max-
tricular function, Analysis of the CASS data has in- imal exercise testing: 10 years’ experience of Seattle Heart
dicated that these patients benefit from coronary ar- Watch. J Am Co11 Cardiol 1983;2:565-73.
tery bypass surgery with respect to survival. An ex- 14. Bruce RA. Exercise testing of patients with coronary heart
disease. Ann Clin Res 1971;3:323-32.
ercise capacity of greater than 10 METS designates 15. Hammermeister KE, DeRouen TA, Dodge HT. Variables
an excellent survival group, again despite the extent predictive of survival in patients with coronary disease.
of coronary artery disease or left ventricular function. Circulation 1979;59:421-30.
16. McNeer JF, Margolis JR, Lee KL, et al. The role of the exer-
If 10 METS truly exerts a “protective effect” that cise test in the evaluation of patients for ischemic heart dis-
obviates any survival benefit from coronary artery ease. Circulation 1978;57:64-70.
bypass surgery, this has enormous implications for 17. Platia EV, Grunwald L, Mellitis ED, et al. Clinical and arte-
riographic variables predictive of survival in coronary artery
cost containment and medical care. It is nonetheless disease. Am J Cardiol 1980;46:543-52.
important to remember thatthis level of exercise ca- 18. Podrid PJ. Graboys TB, Lown B. Prognosis of medically
pacity does not imply the absence of either coronary treated patients with coronary-artery disease with profound
ST-seement denression during exercise testing. N Engl J Med
disease or triple-vessel coronary disease. 1981;%5:1111-6.
Exercise capacity is related to more than just car- 19. Kent K, Rosing DR, Ewels CJ, et al. Prognosis of asymptom-
diovascular fitness and integrity. It is dependent atic or mildly symptomatic patients with coronary artery dis-
ease. Am J Cardiol 1982;49:1823-31:
upon a combination of other physiologic components 20. Sawada SG, Ryan T, Conley MJ, et al. Prognostic value of
as well, including pulmonary function, health status normal exercise echocardiogram. AM HEART J 1990;120:49-55.
of other organ systems, nitrogen balance, nutritional 21. Weiner DA, Ryan TJ, McCabe CH, et al. Prognostic impor-
tance of a clinical profile and exercise test in medically treated
status, medications, orthopedic limitations, and oth- patients with coronary artery disease. J Am Co11 Cardiol
ers. Nonetheless, exercise capacity has been estab- 1984;3:772-9.
lished as a powerful predictor of survival and de- 22. Weiner DA, Ryan TJ, McCabe CH, et al. The role of exercise
Volume 122
Number 5 Prognostic value of exercise capacity

testing in identifying patients with improved survival after painless strongly positive exercise electrocardiogram. Am J
coronary artery bypass. J Am Coll Cardiol 1986;8:741-8. Cardiol 1988;62:892-5.
23. Weiner DA, Ryan TJ, McCabe CH, et al. Value of exercise 26. Bogaty P, Dagenais GR, Cantin B, et al. Prognosis in patients
testing in determining the risk classification and the response with a strongly positive exercise electrocardiogram. Am J
to coronary artery bypass grafting in three-vessel coronary ar- Cardiol 1983641284-8.
tery disease: a report from the Coronary Artery Surgery Study 27. Dagenais GR, Rouleau JR, Christen A, Fabia J. Survival of
(CASS) registery. Am J Cardiol 1987;60:262-6. patients with a strongly positive exercise electrocardiogram.
24. Mark DB, Hlatky MA, Harrell FE, et al. Exercise treadmill Circulation 1982;65:452-6.
score for predicting prognosis in coronary artery disease. Ann 28. Harris PJ, Harrell FE, Lee KL, et al. Survival in medically
Intern Med 1987;106:793-800. treated coronary artery disease. Circulation 1979;60:1259-69.
25. Dagenais GR, Rouleau JR, Hochart P, et al. Survival with

Constrictive pericarditis versus restrictive


cardiomyopathy: A reappraisal and update of
diagnostic criteria
Paul T. Vaitkus, MD, and William G. Kussmaul, MD. Philadelphia, Pa.

The difficulty in distinguishing between constrictive rithm that can be applied when either constrictive or
pericarditis and restrictive cardiomyopathy has been restrictive disease is suspected.
the subject of numerous reports and reviews. Despite
HEMODYNAMIC CRITERIA
the application of a variety of available criteria, some
patient’s conditions still confound attempts at diag- The first recording of intracardiac hemodynamic
nosis. Reviews and standard textbooks of cardiology measurements in constrictive pericarditis were per-
have commented on the limitations of invasive he- formed in 1946 by Bloomfield et al? Subsequently,
modynamic criteria, l-5 but none has quantitatively Hansen et al.s described the classic“dip and plateau”
examined the accuracy of these criteria since the re- or “square root” appearance of the diastolic tracing
port of Wood6 three decades ago. Furthermore, in in constriction. Later studies demonstrated similar
recent years a number of reports have described new recordings in restrictive cardiomyopathy.g Three
methods for distinguishing these two conditions. The well-known hemodynamic criteria for distinguishing
purposes of this review are: (1) to critically examine the two conditions are: (1) Equalization (difference
the hemodynamic criteria applied in the effort to a5 mm Hg) of right ventricular (RV) end-diastolic
separate these two groups of patients; (2) to review pressure (RVEDP) and left ventricular (LV) end-di-
the current status of noninvasive testing and myo- astolic pressure (LVEDP) favors constriction. (2)
cardial biopsy in these conditions; and (3) on the ba- Constriction is more often associated with modest
sis of the data presented, to develop a clinical algo- elevations of RV systolic pressure (550 mm Hg),
whereas in restriction RV systolic pressures fre-
quently exceed 50 mm Hg. (3) Constriction is associ-
ated with RVEDP exceeding one third the RV
From the Cardiovascular Division, Hospital of the University of Pennsyl- systolic pressure, whereas in restriction the ratio is
vania.
characteristically less than one third.’ The first cri-
Received for publication March 18, 1991; accepted May 7, 1991.
terion was proposed by Wood6 and the latter two cri-
Reprint requests: Dr. William Kussmaul, Cardiovascular Division, F.9.121,
Hospital of the University of Pennsylvania, 3400 Spruce St., Philadelphia, teria were first articulated by Yu et al.l”
Pa 19104. Wood6 evaluated the ability of equalization of
4/1/32453 LVEDP and RVEDP to distinguish constriction

1431

You might also like