Download The uses and limitations of end-systolic indexes of left

PERSPECTIVE
The uses and limitations of end-systolic indexes of left
ventricular function
BLASE A. CARABELLO, M.D.,
AND
JAMES F. SPANN, M.D.
Downloaded from http://circ.ahajournals.org/ by guest on April 30, 2017
THE EXTENT of left ventricular dysfunction produced by heart disease is a major determinant of clinical prognosis. Consequently, it is important to assess
left ventricular function accurately. Ejection phase indexes, particularly ejection fraction, are the most
widely used clinical measures of left ventricular function. Ejection fraction is determined easily both by
invasive and noninvasive techniques and, because it
has no dimensions, it is not necessary to correct it for
body size. Unlike isovolumetric indexes such as maximum velocity, ejection fraction does not require extrapolation. These characteristics have made the ejection fraction useful in making clinical decisions
regarding left ventricular function. However, ejection
fraction is affected by changes in loading conditions
independent of changes in muscle function and therefore may not assess the latter accurately in patients
with that alter load. For example, in mitral regurgitation, preload is increased due to the regurgitant lesion
and afterload may be decreased as the left ventricle
empties into the relatively low-pressure left atrium.
These factors maintain a normal ejection fraction even
when significant left ventricular dysfunction is present. Preoperative ejection fraction may fail to predict
the poor or even fatal clinical result of mitral valve
replacement in such patients. In patients with aortic
stenosis, excessive afterload often reduces ejection
fraction when muscle function is relatively normal.
Despite a severe reduction in ejection fraction, these
patients may still benefit from aortic valve replacement. Thus, reliance on ejection fraction can cause
clinical misjudgment regarding the timing and risk of
cardiac surgery. New methods to evaluate left ventricular function that are independent of, or account for,
loading conditions are currently being evaluated and
are reviewed below. These methods, described by
From the Section of Cardiology, Department of Medicine, Temple
University Hospital, Philadelphia.
Address for correspondence: Blase A. Carabello. M.D., Section of
Cardiology, Temple University Hospital, Philadelphia, PA 19140.
Received Sept. 29, 1983; revision accepted Jan. 4, 1984.
1058
Sagawa in a well-referenced review article and editorial' and by others, rely upon known physiologic principles of the left ventricle at end-systole.
Briefly, as shown in figure 1, the volume remaining
in the ventricle at end-systole is not dependent on the
initial ventricular volume. In practical terms, this
means that end-systolic volume is independent of preload. It has therefore been suggested that end-systolic
volume index or dimension might be a better indicator
of contractile function than ejection fraction in states in
which preload is greatly altered.
End-systolic dimension and end-systolic volume
index
Preload is increased in mitral and aortic regurgitation. A large end-systolic volume in patients with these
diseases might indicate left ventricular dysfunction
that has made the ventricle unable to empty properly
despite a normal ejection fraction. Henry et al.2 determined left ventricular end-systolic dimension echocardiographically in the evaluation of patients with aortic
insufficiency. In asymptomatic patients, an end-systolic dimension of greater than 55 mm predicted the onset
of symptoms within an average of 30 months if the
aortic valve was not replaced. In symptomatic patients, an end-systolic dimension of greater than 55
mm predicted persistent symptoms and left ventricular
dysfunction even after aortic valve replacement. In
mitral and aortic regurgitation, Borow et al.3 found that
preoperative end-systolic volume, when indexed for
body surface area, was a better predictor of clinical
outcome and postoperative shortening fraction than
preoperative ejection fraction in patients with these
diseases.
Limitations. Both afterload and contractile function,
as opposed to contractile function alone, determine the
end-systolic volume or dimension. Thus, patients with
high end-systolic afterloads (recently reported in some
patients with aortic insufficiency) may have large endsystolic volumes along with relatively good muscle
function. This may explain recent observations that
some patients with aortic insufficiency have a good
CIRCULATION
PERSPECTIVE
E
E
LU
501
z
U#
>
0-iT
VENTRICULAR VOLUME mI
FIGURE 1. Pressure-volume loops are shown for an isolated canine
ventricle. End-systolic volume is constant despite a wide range of enddiastolic volumes as long as contractile state and afterload remain constant. (Reprinted with permission from Suga et al., Circ Res 32: 314,
1973.)
Downloaded from http://circ.ahajournals.org/ by guest on April 30, 2017
prognosis despite increased end-systolic dimension. It
may therefore be necessary to "correct" end-systolic
volume for end-systolic afterload to properly gauge
contractile function. This concept has led to the investigation of methods that take afterload into account.
The slope of the end-systolic afterload-end-systolic
volume (or dimension) relationships
While independent of preload, the completeness of
ventricular emptying is dependent on contractile function and the afterload resisting further ventricular emptying at the end of systole. End-systolic afterload is the
counterforce that halts shortening when the afterload
equals the force generated by the muscle. At a given
contractile state, when afterload increases, the ventricle empties less completely and end-systolic volume is
greater. As shown in figure 2, the relationship between
end-systolic force (or afterload) and end-systolic
length (or volume) is linear. ' Afterload is quantified by
force on the Y axis and length is shown on the X axis.
Different cycles recorded at different end-systolic afterloads are shown. The slope of the line relating endsystolic force and length, or pressure and volume,
which can be determined by solving the equation Y =
mX + B for m, is a measure of contractile function.
Increases in inotropic state result in a steepening of this
slope. Thus, at an increased inotropic state the ventricle can shorten to a smaller end-systolic volume or
length at the same or even higher end-systolic afterload
(pressure). Simply put, stronger ventricles shorten
more against higher loads than weaker ventricles. It
has been shown that the end-systolic pressure-volume
or pressure-dimension relationship reflects known
changes in intropic state. ' The slope of the end-systolic
pressure-volume or pressure-dimension relationship
has been useful in evaluating contractile function clinically. For example, Grossman et al.4 demonstrated the
pressure-volume relationship to be reflective of a clinically apparent decrease in contractile function in patients with cardiomyopathy.
In clinical use, the end-systolic volume or end-sys-
2.2
0
FIGURE 2. Force-length loops for a left ventricle
are shown at different end-systolic afterloads. As
afterload is decreased, emptying is more complete,
and end-systolic length is less. The relationship between end-systolic afterload (force or pressure) and
end-systolic length is demonstrated to be linear. (Reprinted with permission from Weber and Janicki.
Am J Cardiol 40: 740, 1977.)
W~~~~~~~~~~~
w0
LL
0.0
15.0
16.3
17.5
LENGTH cm
Vol. 69, No. 5, May 1984
1059
CARABELLO and SPANN
Downloaded from http://circ.ahajournals.org/ by guest on April 30, 2017
tolic dimension is determined by contrast or nuclear
angiography or echocardiography and end-systolic
pressure by catheter or cuff at different loading conditions. Pressure is usually altered by infusion of a
noninotropic vasoconstrictor or vasodilator. The endsystolic pressure-volume or pressure-dimension relationship is obtained at each level of afterload. The X,
Y coordinates obtained in this manner are used to plot
the pressure-volume line from which the slope is derived. Since intraventricular pressure varies only modestly from patient to patient, and intraventricular volume varies significantly with body size, some
investigators have indexed end-systolic volume for
body size. In comparing the contractile state of a normal 50 kg woman with that of a normal 100 kg man, if
a 10% change in end-systolic pressure effects a 10%
change in end-systolic volume in both patients, the
change in pressure will be similar in both patients but
the change in volume in the man will be greater than in
the woman. The slope will be lower in the man even
though contractile state is not different. Indexing for
body size tends to cancel out this difference.
Potential limitations. One potential limitation of the
use of the pressure-volume relationship stems from the
fact that pressure may not be an accurate measure of
end-systolic afterload. Afterload is the force that opposes ejection. In physical terms, force pressure x
area and accounts for the distribution of pressure over
the surface to which the force is applied. When the
force is applied to a thick-chambered sphere, it is more
accurately described by the LaPlace relation stress =
P.r/2h, where P = pressure; r = radius; h = thickness. Since wall thickness and chamber radius are
quite different in different diseases, and even vary from
patient to patient with the same disease, use of wall
stress, which corrects for the variability in wall thickness and chamber radius, offers some advantages. For
example, in our catheterization laboratory we achieved
a marked reduction in the end-systolic volume in a
patient with aortic insufficiency by the infusion of the
vasodilator nitroprusside, but end-systolic pressure did
not change. Thus, the calculated slope of the pressurevolume relationship was 0, a theoretic improbability.
This occurred because a reduction in afterload allowed
greater left ventricular emptying and greater cardiac
output at the same blood pressure. Although pressure
did not change, radius decreased, producing a decrease
in calculated wall stress. In such instances, end-systolic stress, which takes into account changes in left ventricular size as well as pressure, is more useful in
demonstrating reduced afterload. Thus, the end-systolic stress-end-systolic volume relationship may be a
=
1060
better reflection of the slope of the afterload-volume
relationship than is the end-systolic pressure-end systolic volume relationship. The stress-volume relationship has also been used successfully in the assessment
of left ventricular contractile function. In our laboratory, for example, noninvasive evaluation of the stressvolume relationship in patients with sickle cell anemia
revealed contractile depression not evidenced by shortening fraction.5 In a recent study of patients with aortic
stenosis, a reduced slope of the stress-volume relationship identified patients with reduced contractile function. The reduced slope correlated well with clinical
evidence of heart failure.6
Limitations. Changes in the X intercept of the stressvolume (or stress-dimension) line may occur with
changes in inotropic state instead of a change in slope.4
This produces a parallel shift in the stress-volume line.
The explanation for this parallel shift is not known. A
parallel shift is likely to occur when inotropic state and
loading conditions are changed in a small ventricle; in
a small ventricle, a given change in pressure causes
less change in stress than the same change in pressure
in a large ventricle. Since slope = change in Y (pressure or stress)/change in X (volume), slope will change
relatively less in a small ventricle when stress is used
instead of pressure. Thus, the slope of the stress-volume relationship may fail to increase when inotropic
state is increased and a parallel shift may occur.
For enlarged ventricles the use of stress-volume relationships may have an advantage over that of pressure-volume relationships because a change in load on
the ventricle is more likely to be due to change in size
of the ventricle rather than change in pressure. Such a
change in load is better quantified by stress. However,
in smaller ventricles, the use of stress-volume relationships may have a disadvantage over pressure-volume
relationships. This occurs as a result of changes in
stress, which are blunted in comparison with changes
in pressure because the small radius present reduces
calculated stress. In this situation, afterload is better
quantified as pressure.
It also must be pointed out that stress is calculated
value and its determination is subject to some error.
However, calculated stress has been validated experimentally and remains a reasonable estimate of afterload.
Finally, an additional limitation might be imposed
because reflex changes in heart rate and inotropic state
could occur during load-altering maneuvers. However, the linearity of the stress-volume relationship reported in most studies suggests that contractile function does not change very much during loading.
CIRCULATION
PERSPECTIVE
Downloaded from http://circ.ahajournals.org/ by guest on April 30, 2017
Dead volume
The X intercept of the afterload-volume relationship
is the volume that would be left in the ventricle if there
were no load on the ventricle at end-systole. This volume has been termed dead volume or VO. The usefulness of VO as an indicator of inotropic state remains
controversial since it is a theoretic value that cannot be
actually measured. It is extrapolated from the pressurevolume or stress-volume line.
In theory, an increase in inotropic state could decrease VO because the unloaded ventricle could shorten
more when inotropic state is increased. Conversely,
VO could increase when inotropic state diminishes.
Since both pressure and stress are zero at VO (or dead
dimension), it should not matter whether the pressurevolume or stress-volume relationship is used to extrapolate this value. However, this does not seem to be
the case. Marsh et al.7 found marked differences in
dead dimension calculated with a pressure-dimension
extrapolation and that calculated with a stress-dimension extrapolation.
Some authors have found VO useful in measuring
inotropic state. Grossman et al.4 found both VO extrapolated from the pressure-volume relationship and
zero circumference (CO) extrapolated from the stresscircumference relationship to be reflective of contractile state. Greater VO or CO indicated poorer contractile
function. Conversely, Mehmel et al.8 found no consistent relationship between inotropic state and VO extrapolated from pressure-volume relationships. Since
the value is only a theoretic one and cannot be verified
clinically, this index of contractile function may be of
limited use.
Attempts at simplification
Substitution of peak systolic pressure for end-systolic
pressure. An attractive feature of pressure-dimension
and stress-dimension relationships is their potential for
being determined noninvasively. End-systolic dimension and thickness can be determined with echocardiography and blood pressure by sphygmomanometry,
but end-systolic pressure is not the same as either peak
systolic pressure or diastolic pressure, which are the
parameters normally measured by sphygmomanometry. In some studies, however, peak systolic pressure
has been substituted for end-systolic pressure. In one
investigation Marsh et al.7 found a close correlation (r
= .99) between slope of the pressure-volume relationships obtained using peak pressure vs the slope obtained during end-systolic pressure. However, the
close relationship between peak and end-systolic pressure may not be constant. It is conceivable that when
Vol. 69, No. 5,
May
1984
large stroke volumes and pulse pressures are present
(e.g., in aortic regurgitation) the relationship of peak
systolic pressure to end-systolic pressure might be different than the relationship when a small stroke volume
and narrow pulse pressure (e.g., in severe heart failure) are present. In such situations, substitutions of
peak systolic pressure for end-systolic pressure might
not be justified. Marsh et al. have described a method
whereby a noninvasively determined carotid pulse
tracing can be calibrated to the patient's blood pressure
as determined by sphygmomanography. The dicrotic
notch pressure so obtained is considered to be endsystolic pressure. Further use of this technique and
examination of the peak systolic pressure-end-systolic
pressure relationship over a wide range of diseases will
be needed to ascertain which is most applicable noninvasively.
The end-systolic pressure-end-systolic volume ratio. As
noted, the determination of the pressure-volume relationship requires derivation of the pressure-volume or
stress-volume relationship at different loading conditions to establish the slope of the line describing this
relationship. Thus, an intervention must be performed
and a measurement of ventricular size must be made at
each load. This is cumbersome and, in some patients,
hazardous. To avoid these problems, the simple ratio
of pressure and volume at end-systole has been used as
a measure of contractile function. If the pressure-volume line runs through or near the 0 origin of both X
and Y, then any single pressure-volume ratio determination defines the pressure-volume slope m = Y/(X +
0). Since the Y intercept of the pressure-volume line
varies from patient to patient, this assumption is usually not justified and thus the pressure-volume ratio does
not truly represent the pressure-volume slope.
However, a rationale for the use of the pressurevolume ratio does exist. The end-systolic volume or
dimension by itself is a useful indicator of prognosis in
some patients.2 Correction for the changes of endsystolic volume produced by variations in ventricular
end-systolic pressure may make this measurement
even more useful. The use of the pressure-volume ratio
requires the assumption that at higher pressures there
are larger volumes at the same inotropic state. A smaller volume for a given pressure suggests more complete
contraction and should indicate that the inotropic state
has increased. The clinical use of the pressure-volume
ratio is currently controversial; while some authors
have found the simple pressure-volume ratio useful in
segregating patients with left ventricular dysfunction
from those without dysfunction, others have not.
End-systolic stress-end-systolic volume ratio. As noted
1061
CARABELLO and SPANN
MITRAL REGURGITATIO N
EJECTION FRACTION
END SYSTOLIC INDEX
7.0-
00
0
0.8-
A
o
6.0-
0
0.75.0-
0
0
0
a
80
0
0
0)
4.0-
0
".
0
0.6-
0*
a
0 0
0
S0
0
3.00.5-
A
*
2.0-
A
Downloaded from http://circ.ahajournals.org/ by guest on April 30, 2017
0.4-
A
I AUl.
FIGURE 3. Left, The end-systolic stress-end-systolic volume index ratio for normal subjects (open circles), patients with mitral
regurgitation who improved clinically after valve replacement (solid circles), and patients with mitral regurgitation who died or
did not improve with surgery (open triangles). The ratio allowed separation of these groups of patients while use of ejection
fraction (right) did not. (Reprinted with permission from Carabello et a.')
above, the use of data on stress often has advantages
over use of that on pressure in quantitating afterload
when ventricular size and thickness are abnormal. The
use of end-systolic wall stress instead of pressure is
thus a better way to correct end-systolic volume for
afterload. Therefore, the end-systolic stress-volume
ratio should be more useful than the pressure-volume
ratio. The determination of the end-systolic stress-volume ratio does not require load manipulation and thus
is simple and safe and the ratio is useful because
whether an increase in inotropic state produces an increase in slope of the stress-volume line or a decrease
in the X intercept (parallel shift to the left), the endsystolic volume for any given end-systolic stress will
be less. Thus, the stress-volume ratio will increase in
either case and properly indicate the positive inotropic
change. Also, a reduction in inotropic state will result
in a larger end-systolic volume for any level of endsystolic stress and the ratio will decrease, properly
indicating reduced inotropic state. When applied clinically, the end-systolic stress-end-systolic volume index ratio separated patients with a good prognosis for
valve replacement for mitral regurgitation from those
with a poor prognosis .9 The ratio was useful despite the
difficulty in defining end-systole in mitral regurgitation, in which volume may still be ejected into the left
atrium in early diastole. Four of five patients with a
reduced ratio died in the perioperative period and the
fifth was clinically unimproved after surgery. As
1062
shown in figure 3, this ratio was a better predictor of
clinical outcome than ejection fraction. We have also
demonstrated a reduction in the end-systolic stressvolume index ratio in patients with aortic regurgitation
and congestive heart failure and in patients with sickle
cell anemia, who also had a reduced stress-volume
slope. Thus, in these studies, the ratio of end-systolic
stress to end-systolic volume index has clinical usefulness. However, it is likely that limitations will be noted since this ratio, like the pressure-volume ratio, is
not a substitute for the stress-volume slope. The ratio
might be particularly limited in very small ventricles.
The formula for the straight line describing the endsystolic relationship is stress - m(ESV -V0), where
ESV is end-systolic volume. The smaller the ventricle
at end-systole, the larger V0 becomes in relation to
end-systolic volume, and the more important V0 becomes in the equation.
Shortening fraction-stress and ejection fraction-stress
relationships. Several authorsv have noted an inverse
correlation between ejection fraction and end-systolic
stress. As end-systolic stress rises, ejection is limited
by the increased load and ejection fraction decreases.
This relationship is linear. A decrease in contractile
function is suggested when ejection fraction, at any
given stress, is lower than predicted (figure 4). This
suggests that high afterload alone cannot account for
the observed reduction in ejection fraction and that left
ventricular dysfunction is present. We and others have
CIRCULATION
PERSPECTIVE
H
.6
6.5
I-.
c:
.4
v
\
* \
*-
X
.2
X
X
i*
\
X
Downloaded from http://circ.ahajournals.org/ by guest on April 30, 2017
I
I
150
200
250
a
(dyneo-s
X
3/CI m
103/c' M2t
400
FIGURE 4. The ejection fraction-stress relation iship for patients with
aortic stenosis, congestive heart failure, and a go()d operative result (D)
is shown together with that for similar group of paitients who had a poor
operative result (X). The latter had a lower ejecti( n fraction for a given
afterload, suggesting that muscle dysfunction wais present. (Reprinted
with permission from Carabello et al. `)
used this relationship to examine left ventricular function in patients with a pressure ov(erload and have
found it useful in identifying patients with left ventricular dysfunction. °
Potential limitations. This relationshiip may not hold
true in patients with mitral regurgitattion in whom increased preload might elevate ejectiorn fraction despite
left ventricular dysfunction. In such c;ases, the ejection
fraction-stress relationship might fa il to shift to the
left, concealing diminished contractil e function. In patients with mitral stenosis reduced pre load may reduce
ejection fraction. The ejection fractioyn-afterload relationship might falsely indicate contraLctile dysfunction
when none exists in patients with th is disease.
Our perspective
Historically, clinical indexes of leflt ventricular contractile function have gained acceptanice and then have
fallen into disuse because their limitattions were identified. It is likely this process will con tinue. Currently,
ejection fraction is the most widely applied index of
ventricular function. It should remain clinically useful
in the treatment of those with corornary disease and
other diseases that do not severely alt' er loading conditions. However, in patients with hleart disease, in
Vol. 69, No. 5, May 1984
whom major alterations in loading conditions occur,
ejection fraction is not an accurate measure of contractile function. In these patients a measure of contractile
function that is independent of or accounts for load is
needed to accurately assess ventricular function. Studies by Sagawa and others have demonstrated that the
end-systolic afterload-end-systolic length relationship
is sensitive to changes in contractile state and insensitive to load. We believe that this relationship offers an
important advance for the investigator and for the clinician in the evaluation of ventricular function. It is not
yet clear what the best method is for determining the
end-systolic relationship.
We believe that, for the purposes of clinical investigation, the end-systolic pressure-volume (or dimension) relationship, constructed from several points obtained under different loading conditions, is currently
the most precise way of assessing muscle function.
This relationship has been shown to be sensitive for the
detection of depressed contractile state but insensitive
to load. The method has already advanced our knowledge of muscle function in patients with diseases such
as thalassema, sickle cell anemia, and cardiomyopathy. Stress-volume or stress-dimension relationships
should be used when cardiac hypertrophy is present
since stress is a better estimate of afterload than pressure in this circumstance. However, because of the
difficulties with performing load manipulation in sick
patients, either technique is limited as an aid to the
clinician in daily decision making.
For clinical use, the ratio of end-systolic stress to
end-systolic volume index offers a promising, easily
obtained measure of ventricular function. We believe
it should be particularly useful in estimating prognosis
in patients with valvular regurgitation, in whom loading conditions are abnormal and V0 is probably of
reduced importance. However, before this ratio can
gain wide acceptance, normal values and laboratoryto-laboratory variation must be identified. Furthermore, prospective studies to establish the prognostic
value of the end-systolic stress-volume index ratio are
needed to confirm or refute its usefulness.
We thank Maxine Blob, Sharon Jankowski, and Rosalie Ratkiewicz for preparation of the manuscript.
References
1. Sagawa K: The end-systolic pressure-volume relation of the ventricle: definition, modifications and clinical use. Circulation 63:
1223, 1981
2. Henry WL, Bonow RO, Rosing DR, Epstein SE: Observations on,
the optimum time for operative intervention for aortic regurgitation. II. Serial echocardiographic evaluation of asymptomatic patients. Circulation 61: 484, 1980
3. Borow KM, Green LH, Sloss LJ, Braunwald E, Collins JJ, Cohn
1063
CARABELLO and SPANN
4.
5.
6.
7.
L. Grossman W: End-systolic volume as a predictor of post-operative left ventricular performance in volume overload from valvular
regurgitation. Am J Med 68: 655, 1980
Grossman W, Braunwald E. Mann T. McLaurin LP. Green LH:
Contractile state of the left ventricle in man as evaluated from endsystolic pressure-volume relations. Circulation 56: 845, 1977
Denenberg BS. Criner G. Jones R. Spann JF: Cardiac function in
sickle cell anemia. Am J Cardiol 51: 1674. 1983
Spann JF. Bove AA. Natarajan G. Kreulen T: Ventricular performance, pump function and compensatory mechanisms in aortic stenosis. Circulation 62: 576. 1980
Marsh JD. Green LH. Wynne J. Cohn PF. Grossman W: Left
ventricular end-systolic pressure-dimension and stress-length rela-
tions in normal human subjects. Am J Cardiol 44: 1311, 1979
8. Mehmel HC. Stockins B. Ruffman K, Olhausen K. Schuler G.
Kubler W: The linearity of the end-systolic pressure-volume relationship in man and its sensitivity tor assessment of left ventricular
function. Circulation 63: 1216. 1981
9. Carabello BA, Nolan SP. McGuire LB: Assessment of preoperative left ventricular function in patients with mitral regurgitation:
value of the end-systolic wall stress-end-systolic volume ratio.
Circulation 64: 1212, 1981
10. Carabello BA. Green LH. Grossman W. Cohn LH, Koster JK.
Collins JJ: Hemodynamic determinants of prognosis of aortic valve
replacement in critical aortic stenosis and advanced congestive
heart failure. Circulation 62: 42, 1980
Downloaded from http://circ.ahajournals.org/ by guest on April 30, 2017
1064
CIRCULATION
The uses and limitations of end-systolic indexes of left ventricular function.
B A Carabello and J F Spann
Downloaded from http://circ.ahajournals.org/ by guest on April 30, 2017
Circulation. 1984;69:1058-1064
doi: 10.1161/01.CIR.69.5.1058
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1984 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://circ.ahajournals.org/content/69/5/1058.citation
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/