Download Reproducibility of Echocardiographic Left Ventricular

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Heart failure wikipedia , lookup

Cardiac contractility modulation wikipedia , lookup

Electrocardiography wikipedia , lookup

Hypertrophic cardiomyopathy wikipedia , lookup

Mitral insufficiency wikipedia , lookup

Ventricular fibrillation wikipedia , lookup

Quantium Medical Cardiac Output wikipedia , lookup

Arrhythmogenic right ventricular dysplasia wikipedia , lookup

Transcript
Reproducibility of Echocardiographic
Left Ventricular Measurements
DONALD C. WALLERSON AND RICHARD B. DEVEREUX
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
SUMMARY Serial echocardiograms with acceptable reproducibility of measurements may be produced by
careful performance and interpretation of the studies. The following recommendations have been shown to
enhance reproducibility. 1) Strict adherence to quality control is necessary to generate echocardiograms of the
highest technical quality. Sonographers should be aware of the definition of a technically adequate study —
including correct beam or plane angulation and continuous visualization of interfaces — and seek this ideal in
every study. Participation by the sonographer in performance of measurements enhances recognition of the
requirements for accurate quantitative echocardiography. Regular machine calibration is a prerequisite to
accurate quantitative echocardiography. 2) Considerable effort must be made to standardize the position of each
acoustic window and angulation from which the patient is imaged — with deviation from these norms being
recorded for future reference. If at all possible, measurements should be taken at end expiration. If that is not
possible, measurement of several consecutive beats will limit the impact of respiratory variation. 3) A uniform
convention of measurement should be adopted. The best candidates for M-mode measurements are the American
Society of Echocardiography recommendations for general measurement and the Penn convention for calculation
of M-mode left ventricular mass. Further data is needed to determine which approaches to two-dimensional
measurements best combine accuracy and reproducibility. 4) Interpretation of echocardiograms may be made
most reproducible by measuring pertinent parameters from multiple beats and using the mean as the result and by
having at least two readers interpret each echocardiogram, possibly with two separate readings by each reader.
(Hypertension 9 [Suppl II]: II-6-II-18, 1987)
KEY WORDS • echocardiography
• left ventricle • reproducibility
O
VER the past decade echocardiography has become
accepted as a valuable noninvasive method for evaluating congenital and acquired cardiac disease. The
initial echocardiographic technique, time motion or M-mode
echocardiography, produced images with excellent depth and
temporal resolution (1 mm and 1 msec, respectively). It was
soon recognized, however, that for some patients, the selected
narrow portions of the heart visualized by the M-mode beam
provided incomplete or even misleading information about the
anatomic relationships of cardiac structures or function of the
heart as a whole. More recently, two-dimensional echocardiography, a technique with superior spatial orientation but inferior
temporal resolution (30 msec) compared to M-mode echocardiography, has extended the usefulness of echocardiography
by providing correct representation of cardiac anatomic
relationships.
In view of the ability of echocardiography to measure cardiac
structure and function noninvasively, investigators soon recognized that standardization of methods was necessary both to
enhance reproducibility of measurements and to facilitate their
comparison between laboratories. Several sources of variability
have been recognized and a number of recommendations to
limit this variability have been published over the last 10 years.
To date, however, individual publications have addressed only
one or a few sources of variability in echocardiographic measurements, and no critical review of variability in quantitative
echocardiographic assessment of the left ventricle is available.
Therefore, this review examines variability of echocardiographic measurements in a comprehensive manner, identifying the
sources of variability in echocardiographic measurement, assessing available approaches to limit such variability, and finally making recommendations to enhance reproducibility of
echocardiographic measurements. The following topics are
considered in sequence: echocardiographic quality control,
equipment calibration, technical factors affecting reproducibility of measurements from serial echocardiograms, physiologic
variation of sequential echocardiographic measurements, echocardiographic measurement convention and reproducibility of
results, magnitude of intraobserver and interobserver variability, temporal variability, and reproducibility of two-dimensional echocardiographic measurements.
From the Division of Cardiology, Department of Medicine, The New
York Hospital-Cornell Medical Center, New York, New York.
Supported in part by Grant HL 18323 from the National Heart, Lung, and
Blood Institute.
Address for reprints: Richard B. Devereux, M.D., Box 222, The New
York Hospital-Cornell Medical Center, 525 East 68th Street, New York,
NY 10021.
II-6
REPRODUCIBILITY OF ECHOCARDIOGRAPHIC MEASUREMENTS/Wa//«rjon and Devereux
Echocardiographic Quality Control
Every investigation that employs echocardiography is dependent on the technical quality of the tracings. Very rarely are the
inclusion or exclusion criteria for echocardiograms specified in
published descriptions of methods. Uniformity in the definition
of an adequate echogram is necessary if strictly comparable data
are to be obtained in different laboratories. This is an especially
important factor in cooperative studies or in population surveys.
Schieken et al.1 have precisely defined a technically satisfactory
M-mode echocardiogram of the left ventricle as comprising the
following:
• Generation of a single dominant line representing each interface
being imaged.
• Demonstration of continuous interface lines at least 5 mm in
length at the point of measurement.
• Demonstration of interfaces with the motion pattern characteristic of the specific cardiac structure being imaged.
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
An additional requirement for timing of echocardiographic
measurements, particularly of end-diastolic dimensions, is the
simultaneous recording of QRS complexes with readily identifiable onset and peak of deflections. These criteria were found by
Schieken et al.1 to result in high reproducibility of measurements of the left ventricle, left atrium, aortic root, and left
ventricular (LV) ejection indices. Twenty tracings judged satisfactory by these criteria yielded highly reproducible measurements, whereas measurements from eight echocardiograms not
meeting these criteria were described as nonreproducible.1
For LV mass measurements, analogous criteria were proposed by Devereux and Reichek.2 LV echograms had to demonstrate continuous motion of right and left septal surfaces and
endocardial and epicardial interfaces of the posterior LV wall
throughout the cardiac cycle. Measurements were made on LV
views at the level of the LV minor axis, identified at or just
below the tips of the mitral leaflets. Measurements were made at
the peak of the R wave of the electrocardiogram. LV mass
measurements in 24 tracings deemed technically adequate by
these criteria exhibited good reproducibility between two experienced observers (r = 0.94, p < 0.001).3
Enforcement of criteria for technical adequacy will exclude
tracings from a proportion of patients. The percentage of excluded tracings will decrease to acceptable levels as technicians
and physicians performing the studies become familiar with the
criteria. It is our impression and that of other investigators that
the yield of technically excellent echocardiograms is enhanced
by involving sonographers in the performance of measurements
on tracings generated by them.
Several other variables appear to influence the percentage of
technically adequate tracings. Specific disease states (especially
pulmonary diseases, thoracic deformities, and postsurgical
changes in cardiac position) may dramatically reduce the yield
of satisfactory studies. The proportion of measurable echocardiograms in population studies has been variably reported as 20
of 28 (71%) by Schieken et al., 1 196 of 259 (75%) by Valdez et
al., 4 and 191 of 236 (81%) by Wong etal. 5 In a recent survey of
normotensive and hypertensive members of an adult employed
population, we found LV echograms to be measurable by strict
criteria in 621 of 767 (81%).6
Equipment Calibration
Echocardiographic measurements depend for their accuracy
on correct calibration of the equipment being used. Errors in
calibration should be rare when equipment is delivered from the
factory but develop commonly thereafter, particularly when
machines are heavily used and moved frequently for bedside
examinations or travel between different research sites. On
II-7
some echocardiographs, calibrations can be correct on one output device but not another (i.e., stripchart recorder vs videotape), necessitating regular checks of the accuracy of each method of display.
M-mode echocardiographs, which image in only a single
direction, can be calibrated with the aid of simple blocks of
materials in which the velocity of ultrasound is known. The
distance between opposite sides of the blocks, in different axes,
can then be made to correspond to various numbers of centimeters in the body, using the ratio of transmission of ultrasound
in body tissues and the plastic of the block. Methods for checking two-dimensional calibration are newer and more complicated since lateral resolution problems have an important effect on
the image produced. Calibration equipment must also mimic the
acoustic properties of the soft tissues through which the echocardiograms are produced. Phantoms for calibration of twodimensional echocardiographs, designed with these considerations in mind, are now available and should be used routinely
by all echocardiographic laboratories to check horizontal and
vertical calibrations as well as resolutions in different axes. An
example of the image produced from one of these two-dimensional phantoms is shown in Figure 1. The method allows calibration checks from nylon lines and cylinders set in a tissuemimicking matrix. It also allows an appreciation of the
inferiority of lateral to depth resolution in currently available
machines.
Technical Factors in Reproducibility of
Echocardiographic Measurements
In addition to the visual quality of echocardiographic tracings, other aspects of echocardiographic technique have been
examined with regard to their effect on reproducibility of measurements. This question has been addressed principally by
studies in which serial echograms performed on the same individuals were separated by time intervals too short to permit real
changes to develop in cardiac structures. This study design
delineated aspects of echocardiographic technique that are important to standardize or replicate to maximize reproducibility.
Patient position during echocardiography and transducer location on the body surface greatly influence the echocardiographic image obtained. Clark et al.7 documented the need to
record accurately both the patient and bed position yielding the
best measurable tracings on the initial echocardiographic study.
This information is particularly important in serial studies, in
which these positions must be reproduced. A number of positions may appear adequate but produce modestly different echocardiographic images of the same structure.3' 8 Most laboratories use the partial (30-45 degrees) left lateral decubitus
position for recording M-mode echocardiograms as well as long
and short axis parasternal and apical four-chamber two-dimensional views. In some patients, optimal recording of apical fourand two-chamber views may require a steeper left lateral position, achieved with a mattress from which a segment has been
removed to permit optimal transducer access. A flat supine
position is used for two-dimensional imaging from the subcostal
acoustic window. The degree of left lateral positioning may be
standardizediaalaboratory by using a, wedge {or oae-ot several
wedges of different shapes) placed under the patient's back.
Many patients require varying degrees of repositioning to accomplish adequate imaging. When that occurs, the details of
such repositioning should be a part of the echocardiographic
record. With regard to bed positioning, a 30-degree upright tilt
of the head is most widely used and would be appropriate for
uniform adoption to reduce interlaboratory variability. Any deviation from the laboratory's routine should be a part of the
echocardiographic record.
II-8
ECHOCARDIOGRAPHY
SUPPL D HYPERTENSION, VOL 9, No
2, FEBRUARY
1987
FIGURE 1. Still frame oftwo-dimensional echocardiographic image, made using a
commercially available calibration phantom, illustrates method of checking horizontal and vertical calibration and depth
resolution. Distance markers at edges of
figure are 1 cm apart.
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
Popp et al.8 showed that the resultant variation in LV dimensions depended on whether the echocardiographic beam paralleled the LV minor axis or was angled obliquely to it, resulting
in overestimation of LV internal dimensions and wall thickness.
Figure 2A illustrates the excellent reproducibility of LV enddiastolic dimensions and fractional shortening on different
strips recorded by Popp et al.8 from the same chest wall location, whereas considerable scatter is seen when dimensions are
obtained from other transducer locations (Figure 2B). Calculations made by Evans et al." revealed that angulation errors
The most important technical issue influencing reproducibility of serial echocardiography is transducer position during
imaging.'• 8 Based on the results of early validation studies,9 it
was recommended that the standard transducer position be "in
the third or fourth interspace, left sternal border, to allow simultaneous recording of continuous endocardial echoes from both
the left ventricular posterior wall and the interventricular septum."10 It was soon recognized, however, that LV echograms
satisfying the above criteria could be obtained from multiple
interspaces with corresponding variations in LV dimensions.
ah
-I
SOj
1
FIGURE 2. Upper panel: A. Comparison
of the end-diastolic left ventricular dimension measured with the transducer in the
standard intercostal (D^) space at different times during the same study period
(Djs-l and DjS-2). B. Comparison of the
end-diastolic left ventricular dimension
measured from the standard intercostal
space (Djs) versus that of the next highest
(DJi) and next lowest (DJ) intercostal
space. Lower panel: A. Comparison of the
change in left ventricular dimension from
end diastole to end systole (AD) from the
standard intercostal space (Ds) at different times during one study period (ADs-l
and ADs-2). B. Comparison of the change
in left ventricular dimension from end diastole to end systole (AD) from the standard intercostal space (Ds) compared with
the next highest (D,,) and next lowest (Df)
intercostal spaces. (Reprinted from Popp
et al.8 with permission.)
I
40
4O|
M
H
JO
*°!
10
(mm)
K>
30
30
40
50
60
(mm)
10
«O
SO
40
SO
tO
14
34
JO
14
.3
13
•
4
(mm)
4
•
13
M
30
34
(mm)
B
4
8
It
16
^
«O
11-9
REPRODUCIBILITY OF ECHOCARDIOGRAPHIC MEASUREMENTS/Wa//er.so/i and Devereux
TABLE 1. Magnitude of Errors Associated with Measuring Ejection Fractions and Rates of Circumferential Fiber Shortening
Source
Typical
value
(%)
Ejection Fractions
Unknown shape and mode of contraction of heart
Rotation
Side-to-side movement
±5
Negligible
+ 12
Papillary muscle volume
-3
Beam out of plane of minor axes
+6
Change in shape of ventricle from systole to diastole
Total
MEASUREMENT
TEMPORAL
VARIABILITY
VARIABILITY
±2
Beam off the major axis
Beam at angle to plane of minor axes
REPRODUCIBILITY
COEFFICENT OF VARIATION
-12
±5
EDD
+ 30
2.6%
-27
(S.D./MEAN)
%D
EDD
11.6%
2.9%
%D
9.7%
Rates of Circumferential Fiber Shortening
Side-to-side movement
±1
Beam off the major axis
+4
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
Beam out of plane of minor axes
+2
Beam at angle to plane of minor axes
-4
Change in shape of ventricle from systole to diastole
±2
Time measurement
Total
±5
+ 14
-12
Based on potential errors in single-dimension echocardiographic measurements.
Adapted from Evans et a l . " with permission.
potentially introduce quantitatively more important errors into
ejection phase indices than do side-to-side motion of the heart or
change in LV shape through the cardiac cycle (Table 1).
In some patients the so-called standard interspaces are not
usable for adequate echocardiographic imaging due to the effects of disease states or unusual body habitus. In obese patients
or patients with ascites and elevated diaphragms, the superior
shift of the heart may allow imaging only from the second
intercostal space. Conversely, in an individual with a long trunk
the heart may lie low in the thorax and imaging may be possible
only from the fifth intercostal space.
Thus, patient positioning, bed positioning, and transducer
location must be stable from study to study. A simple approach
to minimizing the variability of these factors is to note patient
angulation, bed position, and interspace as part of the echocardiographic record. Clark et al.7 showed that by adhering to these
simple recommendations, reproducibility was such that serial
measurement of LV end-diastolic dimension producing differences greater than 3 mm could confidently be deemed as a real
change (falling outside 95% confidence limits derived as ± 2
SD; Figure 3). Similarly, differences in LV fractional shortening greater than 5.5% could also confidently be deemed a real
change {see Figure 3),
More elaborate techniques have been proposed to minimize
variability due to body position and transducer location. One
approach adopted by Wong et al. 3 utilized an inclinometer to
assure reproducibility of transducer angulation. Stefadouros and
Canedoe12 developed an elaborate tnangulation device to
achieve the same goal. Both of these approaches resulted in
good reproducibility but prolonged time for performance of the
study and added some difficulty in the use of cumbersome
equipment. In Table 2 the reproducibility of LV measurements
95% CONFIDENCE INTERVAL
(±2 S.D.)
EDD - ±0.3 cm
%D - ±5.5%
FIGURE 3. Evaluation of the reproducibility of serial echocardiograms by Clark et al.7 Measurement and temporal variability in
serial echocardiograms can produce variation in end-diastolic dimension (EDD) of ± 0.3 cm and in the percentage of minor diameter systolic shortening (%D) of ± 5.5%. (ReprintedfromClark et
al.7 with permission of the American Heart Association.)
obtained using these devices is compared to that obtained by
Clark et al.7 and other investigators in otherwise similar shortterm studies in which reproducibility of patient position and
intercostal space were accomplished manually. The modest
increment in reproducibility attained by more elaborate methods does not justify their use in large-scale echocardiographic
studies.
Physiologic Variability and Echocardiographic
Measurement
The heart is a dynamic organ, constantly changing its size and
function in response to variations in heart rate, preload (influenced by body fluid balance and venous tone), and afterload
(influenced by the relationship between blood pressure and LV
geometry). Variations in these parameters occur both in the
course of normal fluctuations in body physiology over time and
in response to exogenous factors. The latter may include
changes in diet, medications, and emotional stress with its effects on a- and 0-adrenergic tone.
DeMaria et al. l3 examined the effect of heart rate on echocardiographic measurements of LV internal dimension (LVID).
Right atrial pacing was used to increase heart rate in increments
of 10 beats/min to a maximum rate of 150 beats/min. They were
able to show that a 2.7% decrease in LVID occurred, in a
virtually linear fashion, for each 10 beats/min increment in heart
rate (Figure 4). On the other hand, Felner et al.14 and Bellenkie15 did not detect any significant change in the LVID in pa-
11-10
TABLE 2.
ECHOCARDIOGRAPHY
SUPPL II HYPERTENSION,, VOL
9, No
2, FEBRUARY
1987
Reproducibility of Echocardiographic Left Ventricular Measurements
Index of reproducibility
n
Reference
Re-performance variability (%)
Coefficient of variation (%)
Subject
characteristics
LVIDd
LVIDs
—
PWT
[VST
LVIDd
LVIDs
PWT
IVST
With special equipment
Wong et al.5
Stefadouros and Canedoe12
12
Normal size LV
1.8
(49 ±5.9)
12
Dilated LV
4.6
(73 ±8.6)
14
14.5
(50 ±6.4)
—
—
46.6
(35 + 8 0)
63
(50 ±6.4)
7.6
(35 ±8.0)
Without special equipment
Clark etal 7
12
Felnerdal. 1 4
10
6
Pollick et al.43
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
Lapidoet al.44
MacMahon et al
M
Males
Females
—
2.9
(61 ±0.4)
3 1
(48 ±1.5)
4.8
(30.9 ±1 7)
5.0
(8.0 + 0.4)
7.3
(81 ±6)
3.1
(45.7 ±1.4)
4.6
(29.2± 1.3)
6.8
(7.3±0.5)
7.9
(78 ±6)
10
1 beat
4.0
5.8
19.8
12.7
—
—
—
—
10
5 beats
2.4
4.6
11.7
5.7
—
—
—
—
15
Normal subjects
and patients
3.9
4.39
77
10.27
—
—
—
—
10
10
Normal subjects
Independent M-mode
2D guided
—
—
—
—
—
—
4
7
10
11
11
10
14
12
20
20
Patients
Independent M-mode
2D guided
—
—
—
—
—
—
—
—
4
7
7
11
12
10
14
10
14
Outpatients
5.9
(68 ±4)
5.2
(58 ±3)
4
Inpattents
3.0
(67 ±2)
1.8
(57±1)
Pietro et al.43
Gordon et al.61
Valvular and coronary
disease
10
2.2
Normal subjects
10
Congestive myopathy
3.1
(31)
(48)
3.5
(67)
4.3
(55)
Mean values ± SD are in parentheses.
LVIDd = end-diastolic left ventricular internal dimension; LVIDs = end-systolic LVID; PWT = posterior wall thickness; IVST - interventricular septal
thickness; LV = left ventricle; SD = standard deviation; 2D = two-dimensional.
ISO
H 130
FIGURE 4. Relationship of left ventricular (LV) end-diastolic dimension (left panel) and end-systolic dimension (right panel) to heart rate. The LV dimension is
expressedfor each subject as a percentage
of the measurement recorded at a heart
rate of 100 beatslmin, which serves as the
baseline of 100%. The percentage of LV
dimensions in each individual subject is
then correlated with heart rate at 10-beat
increments through a range of 50 to 150
beatslmin. (Reprinted from DeMaria et
al.13 with permission.)
S 100
i
»• - 0 0 2 7 K •! 26
f 90
« • 07
I 50
60 70 80 90 OO 110 120 150 140 140
HEART RATE (o»rra«wl«)
50 60 70 80 90 100 IK) 120 130 140 150
HEART RATE (p«r nunuU)
II-ll
REPRODUCIBILITY OF ECHOCARDIOGRAPHIC MEASUREMENTS/UW/erjo/i and Devereux
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
tients in whom naturally occurring variations in heart rate, in the
normal range of 70 to 100 beats/min, occurred between serial
echocardiograms.
The degree to which variation in heart rate needs to be considered an important determinant of the reproducibility of echocardiographic measurements appears to depend on the expected
fluctuations in heart rate among a group of patients.l3 Based on
our own experience in annual reassessments of apparently normal individuals,16 the mean intraindividual difference in heart
rate between examinations is less than 10 beats/min. This suggests that no more than a 2 to 3% variation in LVID would be
induced by fluctuation in heart rate. Even so, this estimate of
introduced variability undoubtedly overstates the magnitude of
the expected effect since reflex factors that increase heart rate
would also tend to enhance venous return to the heart.
The effect of respiratory variation on echocardiographic measurements of LVID was evaluated by Brenner and Waugh17
after it was initially described by Feigenbaum.18 Brenner and
Waugh17 showed a 6% decrease in end-diastolic LVID at endinspiration compared to end-expiration and recommended that
recording be made at end-expiration. An alternative approach,
when patients have difficulty in cooperating with the recording
of an end-expiration strip maneuver, is to obtain long records of
the best LV views. The pertinent measurements are then
obtained as the mean of up to six consecutive beats, usually
including all phases of a respiratory cycle.
Other physiologic variables beyond the control of the echocardiographer, including preload, afterload, and systemic volume status, are especially likely to influence measurements of
LV function.19"22 While neither preload nor circulatory volume
status can be measured noninvasively, it is possible to measure
accurately an index of LV afterload (LV end-systolic meridional
wall stress) by use of echocardiographic measurement with simultaneous determination of cuff blood pressure.23 We have
shown that resting LV systolic function is inversely related to
end-systolic stress in both normotensive subjects24 and hypertensive patients.25 Similarly, the change in LV fractional shortening observed in both short-term26 and long-term16 studies of
normal subjects can be predicted from the change in end-systolic stress. Thus, analysis of the fractional shortening-end-systolic stress relationship provides a method for determining whether
observed changes in LV systolic function can be explained by
alterations in afterload or may be due to alterations in myocardial contractility.
TABLE 3.
Left Ventricular Measurement
Echocardiographic Measurement Convention and
Reproducibility of Results
To obtain echocardiographic measurements that are reproducible between either observers or studies, the method of making the measurement must be defined. The several sets of measurement conventions that have been recommended are based
either on logic or data.2'1- "• M Three of these are particularly
important because of their use in quantitative studies of LV
mass and function (Table 3). The first of these to be introduced
was the National Institutes of Health (NIH) convention.27 Early
studies utilized measurements by this convention to delineate
anatomic features of hypertrophic cardiomyopathy,27'N athletic
LV hypertrophy,30 and systemic hypertension,31 as well as to
assess the relationships between demographic variables and LV
anatomy.32"33 The fact that LV chamber and wall dimensions
were measured at different times in the cardiac cycle, undermining the logical basis of LV mass calculations, as well as the lack
of necropsy correlation data have caused this method to be
superseded for measurement of LV mass and most other variables by two other conventions, whose introduction was based on
carefully collected data of different types.
The Penn convention, devised in 1977, was based on a study
in which LV mass was calculated from echocardiograms examining two alternative assumptions about each of three variables:
LV shape, wall segments to be measured to determine mean
myocardial thickness, and identification of endocardial surfaces.2 These calculations yielded eight alternative echocardiographic estimates of LV muscle mass, which were systematically compared to necropsy LV mass in 34 patients.2 The set of
assumptions that yielded the most accurate echocardiographic
measurements of LV mass was introduced as the Penn convention (Figure 5; Table 4). Penn convention end-diastolic measurements are taken at the peak of the R wave, and the thickness
of endocardial interfaces is excluded from wall thickness measurements. Although development of the Penn convention was
not based on assessment of interobserver and intertest reproducibility, subsequent study has revealed that reproducibility is
acceptably high (Table 5).
The reverse sequence was followed by Sahn et al.28 in developing the M-mode measurement recommendations of the
American Society of Echocardiography (ASE). In formulating
these recommendations, major emphasis was placed on the reproducibility of measurements performed by 76 readers on a set
Conventions
NIH
(Henry et al. 27 )
Penn
(Devereux and Reichek2)
ASE
(Sahn et al. 28 )
Wall thickness measurements
Timing
Onset of P wave
Peak of R wave
Onset of QRS
Interface definition
Center of interface
Exclude endocardial
thickness from wall
Leading edge to
leading edge
Timing
Maximum LV diameter
Peak of R wave
Onset of QRS
Interface definition
Center of interface
Include endocardia]
thicknesses in LVID
Leading edge to
leading edge
LV mass calculation
Troy formula
Cube formula
Cube formula
Validation
None
Autopsy: regression
equation*
Autopsy: regression
equation!
LVID measurements
NIH = National Institutes of Health; ASE = American Society of Echocardiography; LVID
LVM = left ventricular mass. Other abbreviations as in Table 2.
• L V M p g ^ = 1.04 [(IVS + LVID + PWT) 3 - LVID 3 ] - 13.6 g.
tLVM A S E = 0.80 [1.04 (IVS + LVID + PWT) 3 - LVID 3 ] + 0.6 g
left ventricular internal dimension;
11-12
ECHOCARDIOGRAPHY
SUPPL II HYPERTENSION, VOL 9, No 2, FEBRUARY
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
IVST
IVST
LVID
LVID
PWT
PWT
FIGURE 5. The Penn convention for echocardiographic measurement of interventricular septal thickness (IVST), left ventricular internal dimension (LVID), and posterior wall thickness (PWT)
is illustrated in Panel B. The Penn convention excludes right and
left septa! endocardial interface thickness from IVST and excludes
posterior wall endocardial thickness from PWT. Left septal endocardial thickness and posterior wall endocardial thickness are thus
included in LVID by this method. End diastole is defined as peak R
wave. (Reprinted from Devereux and Reichek2 with permission of
the American Heart Association.)
of sample echocardiographic tracings. The most important features of the ASE measurement convention are that dimensions
are taken from the leading edge of one interface to the leading
interface of the next; identification of end diastole was at the
TABLE 4.
1987
onset of the QRS complex for all measurements; and identification of end systole for LV measurement was at the time of the
most posterior displacement of the interventricular septum in
systole (Figure 6).
Subsequent studies have assessed both the validity and reproducibility of measurement by the ASE convention in comparison with angiographic or necropsy reference standards.3' M ' 3 5
Crawford et al.3* compared LV measurements by the ASE rec^
ommendations with the echocardiographic measurement convention previously established in their laboratory using angiographic LV volumes as the reference standard. Volumes derived
using the cube function formula and echocardiographic LV
measurements by the ASE recommendations consistently overestimated volume relative to angiographic determinations, but
the overestimations were systematic and highly reproducible.
The measurement convention previously used in Crawford's
laboratory afforded a better correlation with angiographically
determined volumes. Although both echocardiographic methods overestimated angiographic LV ejection fraction, the ASE
recommendations produced a better correlation with angiographically determined values.
More recently, Woythaler et al.35 and Devereux et al. 3 ' 36
have assessed the validity of LV muscle mass determination
using echocardiographic measurements by the ASE convention.
Both groups found that LV mass calculated by the ASE convention correlated well with necropsy LV mass but systematically
overestimated it.3' 3}~36 This overestimation can be corrected by
a simple regression equation (see Table 4, equation 3). Other
formulas used to estimate LV mass from M-mode and twodimensional echocardiographic measurements37^10 are also
shown in Table 4 and additional two-dimensional measurements used to calculate LV muscle mass are shown in Figure 7.
In view of these findings, it appears that both the ASE and
Penn conventions provide adequate reproducibility of echocardiographic LV measurements and can be used to estimate LV
muscle mass accurately using necropsy-validated regression
equations.2-3 For most purposes, the ASE convention appears
to be preferable because its wide dissemination makes it suitable
for standardizing measurements among clinical and research
laboratories. For research studies in which a principal goal is
evaluation of LV muscle mass, the Penn convention appears to
Formulas for Echocardiographic Measurement of Left Ventricular Mass
Formula
M-mode echocardiography
Muscle cross-sectional area (CSA) =
n [(LVIDd)/2 + PWTd + IVSTd]2 - n (LVIDd/2)
Penn convention: LV mass (g) =
1.04 [(LVIDd + PWTd + IVSTd)3 - (LViDd)3] - 13.6 g
ASE convention: LV mass (g) •=
0.80 [1.04 X (IVSTd + LVIDd + PWTd)3 - (LVIDd)3] + 0.6 g
Source
Gaash et al.37
Devereux and Reichek2
Devereux et al.3
Two-dimensional echocardiography
Simpson's rule method: LV volume =
-. „
ANT
TN3
£""-•> AT + -^— + —rl
o
Area-length method A: LV volume =
5/6 x (Ap) X L
Area-length method B: LV volume = 1.05 X
[
d3
1
f
d3 1 1
v |(b + t)2 2/3 (a + t) + d - 3 ( a + () j - b 2 [2/3 (a + d) - - j p J J
Helak and Reichek39
Helak and Reichek.3' Reichek et
Schiller et al.38
All dimensions are in centimeters. A = short axis area; IVSTd = end-diastolic intraventricular septal thickness; L = longest
epicardial or endocardial length of the left ventricle; N = number of sections; p = papillary muscle level; PWTd = end-diastolic
posterior wall thickness; T = thickness of each section; a, b, d, t in last equation corresponds to the labels in Figure 7 Other abbreviations
as in Table 2.
11-13
REPRODUCIBILITY OF ECHOCARDIOGRAPHIC MEASUREMENTS/HW/erson and Devereux
TABLE 5. Interobserver and Inienest Reproducibility of Penn Convention Measurements
Measurement
Interval
between
studies
n
Correlation
coefficient
Standard
deviation
Mean
difference
Interobserver
Experienced vs experienced
LVIDd
0
60
0.98
1.4 mm
1.4 mm
Experienced vs experienced
PWT
0
60
0.91
0.9 mm
0.8 mm
Experienced vs experienced
LV mass
0
24
0.94
29 g
26 g
Experienced vs inexperienced
LV mass
0
24
0.84
42 g
38 g
Interstudy
Short-term with necropsy confirmation
LV mass
3 Mo
8
0.98
28 g
26 g
Long-term in normal subjects
LV mass
15 Mo
89
0.78
29g
26 g
Abbreviations as in Table 2.
be slightly preferable because it yielded a closer correlation than
the ASE measurements between echocardiographic and necropsy LV muscle mass.3- x
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
Intraobserver and Interobserver Variability
Echocardiographic measurements are also subject to intraobserver or measurement variability (i.e., fluctuation of measurements on the same echocardiogram when it is measured multiple
times). Additionally, in attempts to enhance objectivity and
accuracy of echocardiographic measurements, studies often employ multiple readers to interpret the same echograms, the mean
of the values produced by all readers being accepted as the final
one. This introduces an additional source of variation, called
interobserver variability. This is a compound variable since
there is also concurrent measurement variability. Vignola et
al.41 described 5.2, 10.0, and 16.4% interobserver errors (ex-
FIGURE 6. American Society of Echocardiography recommendations for Mmode measurements. Diastolic measurements are made at the onset of the QRS
complex of the electrocardiogram (EKG);
cavities and walls are measured at the level of the chordae below the mitral valve.
The illustration and the elliptical inserts a,
b.c.d and e show the leading edge method
as well as measurements using the thinnest
continuous echo lines. ARV = right ventricular anterior wall; RV = right ventricle; LV = left ventricle; PLV = posterior
left ventricular wall; S = septum; PPM =
papillary muscle; AMV, PMV = anterior
and posterior mitral valve leaflets; A, B,
C, D, E and F = points of mitral valve
motion; EN = endocardium; EP = epicardium; Ao = aortic root; AV = aortic
valve; LA = left atrium. The extra line in
insert B, which is excludedfrom the septal
measurement, represents a portion of tricuspid valve apparatus. (Reprinted from
Sahn et al.2S with permission of the American Heart Association.)
FIGURE 7. Left ventricle as a truncated ellipsoid. The internal
dimensions used in this study are shown in the short axis (left) and
the long axis (right). Four semiminor axes or radii (b) are shown in
the short axis and two are shown in the long axis. Note that placement of the minor diameter (equivalent to two semiminor axes)
determines the division of semimajor axes. The full semimajor axis
(a) and the truncated semimajor axis (d) appear in the long axis
representation of the left ventricle. (Reprintedfrom Schiller et al.38
with permission of the American Heart Association.)
11-14
ECHOCARDIOGRAPHY
SUPPL II HYPERTENSION, VOL 9, No 2, FEBRUARY
TABLE 6. QuanxitationoflnterobstrverVariancesforEchocardiographic
Measurements
TABLE 7 Relationship of Percentage Uncertainty for Left Ventricular
Dimensions to Measurement Convention
Criterion for measurement (%)
Measurement
(mm)
Standard error
(mm)
Error as
percentage of mean
LVIDd
±2.35
5.2
LVIDs
±2.32
7.5
IVST
±0.80
10.0
LVIDd
16.4
LVIDs
PWT
±1.35
Abbreviations as in Table 2.
Reprinted from V ignola et al 4I with permission.
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
pressed as the mean difference between observers divided by the
average measurement) for measurement of the end-diastolic
LVID, interventricular septal thickness, and LV posterior wall
thickness, respectively (Table 6). Monoson et al.42 obtained a
correlation coefficient of .91 between measurements of both the
interventricular septal and posterior LV wall thickness of the
same echo tracing by two observers. Sahn et al., 28 evaluating
measurements for end-diastolic LV posterior wall thickness,
interventricular septal thickness, and LVID on the same echocardiograms by 76 observers, showed respective percentage
uncertainties of 23.4%, 19.5% and 8.2% when the ASE convention was used for measurement. Percentage uncertainty was
obtained as follows: "the mean and standard deviation for each
measurement on each recording were determined, first combining all measurement criteria. The 95th percentile confidence
ranges were considered to be 1.97 standard deviations. The
percentage uncertainty was the 95th percentile confidence limit
divided by the mean for the measurement times 100. Percentage
uncertainty is normalized by the mean for the measurement,
allowing comparison of the ranges of errors between the echograms which differed in the absolute measurements." Table 7
shows the percentage uncertainty generated by different measurement conventions for end-diastolic interventricular septal
thickness, posterior wall thickness, and septal LVID as well
as end-systolic LVID. The conventions that produced the
smallest percentage uncertainty were selected as the ASE
recommendations.
Notably, measurement of end-diastolic LVID exhibited the
least interobserver variability in each of these studies.M- 4I The
measurements with the greatest interobserver variability included those of the mitral valve E-F slope, LV posterior wall and
septal thickness, and amplitude of posterior wall excursion.
This observation is not surprising since the boundaries of these
structures are often represented by multiple lines rather than
clear, single, continuous lines. Reader experience appears to
diminish, but not eliminate, interobserver variability, as illustrated by the superior performance of readers with at least 2
years of experience in a laboratory with greater than 800 echocardiograms per year in the study of Vignola et al.41 (Table 6).
Clark et al. 7 sought to determine the most advantageous combination of number of observers and number of readings per
observer. They were able to show that two observers, each
reading each echocardiogram twice at different occasions or
three observers, each reading each echocardiogram once, provided the best reproducibility (Figure 8).
As documented earlier in this review, strict quality control,1
standardization of measurement convention,28 and measurement of multiple beats with the mean value being accepted as
the final reported value6-43 all contribute to decreased intraobserver and measurement variability. Adherence to those principles plus adoption of one of the combinations of observers and
readings recommended by Clark et al. 7 will markedly reduce the
magnitude of interpreter-introduced variability. That this is possible was shown by Lapido et al. 44 in their assessment of the
1987
Measurement
Onset
QRS*
Peak
R wave
8.2
11.8
Peak
posterior
wall
motion
15
IVST
19 5
PWT
23.4
Nadir
Smallest
septa!
LV
motion* dimension
14
16
23.8
23
'Adopted for recommendation by the American Society of Echocardiography.
Abbreviations as in Table 2
Adapted from Sahn et al. 28 with permission.
impact of interobserver variation. The above recommendations
were essentially satisfied by the study's protocol, including use
of the ASE's measurement convention. Their intraobserver and
interobserver agreement was excellent (Table 8).
Temporal Variability
Attempts to define the temporal variability of echocardiographic measurements have utilized serial echograms done on
the same patient in a time period considered too short to allow
any real change in the measured parameters. A number of investigations have attempted to define and limit potential sources of
temporal variation.7's-12- "• 16 - ^ M-M-45 These studies were
designed to develop confidence limits to detect as real changes
only those differences that fall outside the established confidence limits. Clark et al.7 defined a coefficient of variation as
"the standard deviation of all the presented data in a given case
divided by the mean of those measurements." Two standard
deviations on either side of the mean were calculated from the
estimates of reproducibility, and the 95% confidence interval
READERS
1
1
READINGS
1
2
FIGURE 8. Analysis of the percentage of responses falling within
the 95% confidence interval for a given echocardiographic measurement when made by one, two, or three readers making readings
on one or two occasions. Each response was the average of three to
five measurements and each reading was made on separate occasions. (Reprinted from Clark et al.7 with permission of the American Heart Association.)
REPRODUCIBILITY OF ECHOCARDIOGRAPHIC MEASUREMENTS/Wa/Zerion and Devereux
TABLE 8
Blind Duplicate Measurements cf 10 Tracings by Three Observers
LVIDs
LVIDd
PWT
Tracing DO
1
2.
3
6
7
8
9
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
10
Maximum
difference
Mean
absolute
difference
C
A
4.8
4.7
5.2
5.2
4.6
4.4
5.0
4.9
4.7
4.6
5.2
5.4
5.2
5.4
4.9
4.9
5.0
5.0
4.5
4.6
4.9
5.0
4.6
4.5
5.1
5.0
5.1
5.2
4.9
4.9
5.1
5.1
47
4.8
3.3
32
3.0
3.0
3.2
30
30
30
2.8
2.7
4.9
5.1
5.2
5.2
47
4.8
4.9
5.0
5.1
5.1
4.5
4.7
4.9
4.9
4.5
4.2
5.1
5.0
5.1
5.1
5.0
5.0
5.1
5.1
46
4.7
2.8
27
3.7
37
3.6
3.6
3.5
3.5
2.9
3 1
B
3.4
3.3
3.0
30
3.2
3.2
3.0
27
2.9
27
2.8
2.6
3.7
3.7
3.6
3.6
3.5
3.5
2.6
25
0.2
0.3
0.1
0.2
0.3
A
B
C
A
B
3.3
3.3
3.0
3.0
3.2
3.2
3.0
3.1
2.8
2.8
2.8
2.7
3.7
3.7
0.9
0.9
0.9
0.9
1.0
0.8
0.9
0.9
0.8
0.9
1.0
0.9
0.9
0.9
0.8
0.9
1.0
1.0
1.0
1.0
0.8
0.9
0.9
0.8
1.0
0.9
1.0
0.9
.6
.6
.1
.0
3.7
3.7
3.5
3.5
27
2.7
0.9
0.9
0.8
0.8
0.8
0.8
0.9
10
08
0,8
1.0
1.0
1.0
1.0
0.7
08
1.0
09
0.9
09
0.9
0,9
.3
,3
.1
.0
.2
2
.2
.3
0 1
0.1
0.2
0.1
() 3
0.2
0.9
0.9
0.9
1.0
1.0
1.0
1.0
0.9
0.9
0.9
0.9
0.9
10
1.0
IVST
PWT-A
C
.6
.7
.0
.3
OO OO
5
B
OO OO
4
A
.3
.3
.3
.3
.2
.3
.3
.1
.1
.3
.1
.2
.3
.4
.0
(
U-15
C
0.9
1 0
15
1.5
1.0
0.9
1.9
1.6
1.3
1.3
1.3
13
1.1
1 2
1 1
09
A
B
1 4
1.3
0.9
0.9
0.6
0.7
1.1
1.0
1.0
0.9
0.9
0.9
0.9
1.0
0.9
0.8
0.7
0.7
1.1
1.2
0.6
0.6
1.0
1.0
0.6
0.7
1.1
0.9
1.0
1.0
0.9
09
1.0
0.9
1.1
1.1
0.8
0.8
1.2
1.1
0.7
0.7
0.3
0 1
0.1
IVST-A
C
0.9
1.0
0.6
0.7
0.9
1.0
1.0
0.9
1.0
1.0
1.0
1.0
1.1
0.9
0.8
0.8
1.2
1.1
0.7
0.6
0.1
A
B
C
0.7
0.6
0.6
0.6
0.6
0.5
0.6
0.5
1.2
1.2
1.4
1.4
0.7
0.7
0.5
0.5
0.6
0.6
0.8
0.8
0.8
0.8
0.6
0.7
0.6
0.6
0.6
06
1.1
1.1
1.3
1.3
0.6
0.6
0.5
0.5
0.6
0.6
0.9
0.8
0.8
0.7
0.6
0.6
0.7
0.5
0.6
0.6
1.1
1,0
1.4
1.4
0.7
0.7
0.5
05
0.7
07
0.9
0.9
0.1
0.1
0.1
0 07 0.08 0.02
0.03 0.02 0.04
(),08 0.06 0.1
0.06 0.04 0.08
0.12 0.08 0.07
0.03 0.04 0.04
PWT-A = amplitude of posterior wall motion in systole; IVST-A == amplitude of intraventricular septa; motion during systole; other abbreviation.> as in
Table 2.
Adapted from Lapido etal. 44
was determined for each echocardiographic measurement (see
Figure 3). Lapido et al.44 used a similar method to assess temporal variability. Pietro et al.45 expressed temporal variation
(termed re-performance variability) as a "percentage obtained
by the absolute difference between measurements in study 1 and
study 2 divided by the measurement in study 1."
Clarke et al.7 determined that by their approach a change in
end-diastolic LVID of 0.3 cm or greater represented a biologically significant difference. Similarly, a change of 5.5% or
greater in percentage of minor diameter shortening in systole
was biologically significant (see Figure 3). The two studies
noted above obtained acceptable levels of temporal variability
(overall reproducibility). In contrast to those studies, Monoson
et al.42 reported very poor correlations between two sets of
readings by the same observer on echocardiograms performed
on the same patient on different days. It is not clear why this
group obtained such poor reproducibility, but this study stands
out as an exception compared to other studies that have addressed this issue and to our own experience.
One application of echocardiographic methods is the evaluation of LV mass in hypertension.25'3I> ** As is indicated in Table
2, moderately good reproducibility of LV muscle mass determinations have been obtained in serial echocardiograms on the
same subject in our laboratory. An important factor in these
measurements is that the echocardiograms on each subject were
measured totally independently of each other, maximizing the
potential for variability. From data recently reported by other
investigators,47' ** we anticipate that paired reading of serial
tracings on the same subject would reduce this variability by
approximately 50%. This approach introduces the possibility of
systematic observer bias, however, when any clue is available
as to the circumstances under which recordings are taken. Particularly striking examples occur in studies to determine the
effects of )3-blockade or heart valve replacement on LV mass or
function, where the presence of a reduced heart rate or a prosthetic heart valve would indicate which study followed therapeutic intervention.
With the introduction of two-dimensional echocardiographic
imaging, many echocardiographers have concluded, in advance
of any data, that two-dimensional guidance of M-mode echocardiograms would result in greater interstudy reproducibility of
measurements. This was examined by Pietro et al., 45 who were
unable to detect any benefit from two-dimensional guidance.
For example, in their patient population the re-performance
variability for all structures measured was 8.7 ± 0.9% and
9.4 ± 0.7% for independent M-mode and two-dimensional
guided M-mode studies, respectively (Figure 9). A similar lack
of any clear advantage obtained by two-dimensional guidance
was reported by Panidis et al.49
Reproducibility of Two-Dimensional
Echocardiographic Measurements
Two-dimensional echocardiographic imaging has been introduced more recently than M-mode echocardiography, and
hence its quantitative applications have been less extensively
studied. It provides the advantage of a second dimension in
which to view cardiac structures, thus eliminating some errors
in ultrasound beam angulation that otherwise might go unrecognized. But two-dimensional echocardiography also has some
inherent, unique limitations that affect reproducibility. Difficulties common to two-dimensional and M-mode echocardiography include measurement or intraobserver variability, interobserver variability, and temporal variability. Similarly, the
necessity for strict quality control and the importance of standardization of technique are pertinent considerations. A com-
11-16
ECHOCARDIOGRAPHY
SUPPL II HYPERTENSION, VOL 9, No
NORMAL SUBJECTS (n = 1O)
855 20M-Mooe
Ao
LA
LVs
LVo
IVS
I
P»
ABNORMAL SUBJECTS In=20)
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
LA
LVs
LH>
IVS
PW
FIGURE 9. Performance variability was related to specific structure measured but did not differ systematically between independent
and two-dimensional guided M-mode techniques. The left ventricle
diastolic dimension (LVd) is the least variable structure in normal
and abnormal groups. Ao = aortic root; LA = left atrium; LVs =
left ventricle systolic dimension; IVS = interventricular septum;
PW = posterior wall. (ReprintedfromPietro et al.45 with permission.)
mittee of the ASE has published clear recommendations concerning standard transducer positions and imaging planes in an
attempt to standardize the views from which measurements are
made.50 It must be recognized, however, that the two-dimensional tomographic planes may be off axis or misaligned in the
third direction, as shown most clearly by Erbel et al." in a study
in which echocardiographic views from the so-called apical
window were shown to miss the true LV apex (as defined by
angiography) by a wide margin in many patients. Delineation of
technical precautions to prevent this and comparable errors is
urgently needed.
Experience with two-dimensional echocardiographic imaging has identified some internal checks, available to the sonographer, which will limit the error introduced by lack of knowledge regarding the third dimension in the planar images
produced. For example, in determining the correct plane for a
short axis LV view, rigorously producing the smallest circular
cross section at a standardized level (just below the tips of the
mitral valve or at the level of the papillary muscles) will minimize errors of oblique angulation of the beam. A second useful
internal check pertains to the apical four- and two-chamber
views. The true long axis of the left ventricle may be checked by
producing what appears in a four-chamber view to be a maximal
long axis and transverse dimension and then turning the transducer 90 degrees to record the two-chamber view. If the measured long axis is longer in this view, then one can be certain
that the initial four-chamber plane did not maximize the LV
2, FEBRUARY
1987
long axis. In all long-axis or apical views, it is also important to
maximize the transverse dimension of the left ventricle (or other
chamber). Otherwise, the oblique planes may understate chamber size and overstate wall thickness and motion.
Important quantitative applications of two-dimensional echocardiography include measurement of valve areas and LV muscle mass, both of which are supported by careful' validation studies with comparison to appropriate reference standards.38' "*>•32"57 LV wall motion and volumes in asymmetric
chambers can also be assessed by two-dimensional echocardiography although results of numerous proposed approaches are
not entirely consistent.
LV volume and ejection fraction estimation are probably the
most frequent measurements for which two-dimensional echocardiography is used since, theoretically, this mode of imaging
should be superior to M-modein that asymmetric chambers may
be more accurately represented. Reproducibility and accuracy
share a common limitation with M-mode echocardiography,
however: endocardial surface identification. Manipulation of
gain settings may help, but even in the newer generation of
equipment, endocardial surface identification is a difficult task.
Compounding this problem is the observation'in the careful
studies of Helak and Reichek39 that considerable overestimation
of interface thickness was produced by a variety of two-dimensional echocardiographs. This overestimation appeared to be
less prominent in a study by Wyatt et al.58 Efforts to enhance
interface recognition have generated considerable literature.
Computer-assisted endocardial surface identification59 and approximating the LV outline using a calibrated ellipsoid to obviate the painstaking tracing of the endocardial surface outline60
have both been recommended. For serial studies it has also been
proposed to record and replicate gain settings to alleviate variability in interface thickness and location. Although reasonable
correlations between angiographic and two-dimensional echocardiographic determinations of LV volumes have been obtained, two-dimensional echocardiographic methods have systematically underestimated angiographic LV volumes.
Despite these problems, acceptable reproducibility of LV
volume measurements have been obtained by newer two-dimensional echocardiographic methods. For example, Gordon et
al.61 showed that for group data (30 subjects), two-dimensional
echocardiographic determination showing changes of 2% for
LV end-diastolic volume and ejection fraction and 5% for LV
end-systolic volume would be significant. For serial measurements in an individual patient, however, one would need
changes of at least 15% for LV end-diastolic volume, 25% for
LV end-systolic volume, and 10% for ejection fraction to be
confident that the observed changes exceeded 95% confidence
limits.
These data suggest that two-dimensional echocardiographic
LV measurements are sufficiently reproducible to be utilized in
hypertension and epidemiologic research. At the same time,
however, the limited availability of data concerning variability
between serial two-dimensional echocardiograms on the same
subject (as opposed to the more extensive information substantiating reasonable interobserver and intraobserver variability of
measurements of the same echocardiogram62) needs to be corrected. In addition, the reproducibility of two-dimensional measurements needs to be assessed in patients with specific disease
states, as has been done for M-mode measurements in patients
with aortic regurgitation47 or congestive cardiomyopathy.63' M
Acknowledgment
We thank Miss Virginia Bums for her assistance in preparation of this
manuscript.
REPRODUCIBILITY OF ECHOCARDIOGRAPHIC MEASUREMENTS/Wa//erjon and Devereux
References
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
1. Schieken RM, Clarke WR, Mahoney LT, Lauer RM Measurement
criteria for group echocardiographic studies. Am J Epidemiol 1979;
110:504-514
2. Devereux RB, Reichek N. Echocardiographic determination of left ventricular mass ui man: anatomic validation of the method. Circulation
1977,55:613-618
3. Devereux RB, Alonso DR, Lutas EM, et aj. Echocardiographic assessment of left ventricular hypertrophy: comparison with necropsy findings. Am J Cardiol 1986^7:450-458
4. Valdez AS, Motta JA, London E, et al. Evaluation of the echocardiogram as an epidemiologic tool in an asymptomatic population. Circulation 1979;60:921-929
5. Wong M, Shah PM, Taylor RD. Reproducibility of left ventricular
internal dimensions with M-mode echocardiography: effects of heart
size, body position and transducer angulation Am J Cardiol 1981 ;47:
1068-1074
6. Hammond IW, Devereux RB, Alderman MH, etal. The prevalence and
correlates of echocardiographic left ventricular hypertrophy among employed patients with uncomplicated hypertension. J Am Coll Cardiol
1986;7:639-650
7. Clark RD, Viorcuska K, Cohn K. Serial echocardiographic evaluation
of left ventricular function in valvular disease, including reproducibility
guidelines for serial studies. Circulation 1980;62:564-575
8. Popp RL, Filly K, Owen RB, Harrison DC. Effect of transducer placement on echocardiographic measurement of left ventricular dimensions.
Am J Cardiol 1975;35:537-54O
9. Popp RL, Wolfe SB, Hiram T, Feigenbaum H. Estimation of right and
left ventricular size by ultrasound. Am J Cardiol 1969;24:523-530
10. Feigenbaum H. Echocardiography. 1st ed. Philadelphia. Lea and Febiger, 1972:26-28, 99-128
11. Evans DH, McDicken WN, Robertson AR The accuracy of cardiac
function indices derived from ultrasonic time-position scans Cardiovasc Res 1976; 10:65-73
12. Stefadouros MA, Canedoe Ml. Reproducibility of echocardiographic
estimates of left ventricular dimensions. Br Heart J 1977;39:39O-398
13 DeMaria AN, Neumann A, SchubartPJ, Lee G, Mason DT. Systematic
correlation of cardiac chamber size and ventricular performance determined with echocardiography and alterations in heart rate in normal
persons. Am J Cardiol 1979,43.1-9
14. FelnerJM, BlumensteinBA, SchlantRCetal. Sources of variability in
echocardiographic measurements. Am J Cardiol 1980;45:995-1004
15. Belenkiel Beat to beat variability of echocardiographic measurements
of left ventricular end-diastolic diameter and performance JCU 1979;7.
263-268
16. Devereux RB, Hammond IW, Lutas EM, Spitzer MC, Alderman MH,
Laragh JH. Year to year variability of echocardiographic measurements
in normal subjects [Abstract] J Am Coll Cardiol 19843:516
17. Brenner JI, Waugh RA. Effect of phasic respiration on left ventricular
dimension and performance in a normal population, an echocardiographic study. Circulation 1978;57:122—127
18. Feigenbaum H. Echocardiography. 2nd ed Philadelphia. Lea and Febiger, 1976.310-312
19. Burggraf GW, Parker JO. Left ventricular volume changes after amyl
nitrite and nitroglycerin in man measured by ultrasound. Circulation
1974;49:136-143
20 Hirschleifer J, Crawford M, O'Rourke RA, Karliner JS. Influence of
acute alterations in heart rate and systemic blood pressure on echocardiographic measures of left ventricular performance in normal human
subjects. Circulation 1975^2:835-841
21. Redwood DR, Henry WL, Epstein SE. Evaluation of the ability of
echocardiography to measure acute alterations in left ventricular volume. Circulation 1974^0:901-904
22 Nixon JV, Murray RG, Leonard PD, Mitchell JH, Blomqvist CG.
Effect of larger variations in preload on left ventricular performance
characteristics in normal subjects. Circulation 1982;65698-703
23. Reichek N, Wilson J, St. John Sutton M, Plappert TA, Goldberg S,
Hirshfeld JW. Noninvasive determination of left ventricular end-systolic stress: validation of the method and initial application. Circulation
1982;65:99-108
24. Lutas EM, Devereux RB, Reis G, et al. Increased cardiac performance
in mild essential hypertension: left ventricular mechanics. Hypertension
1985;7:979-988
25. Devereux RB, Savage DD, Sachs I, Laragh JH. Relation of hemodynamic load to left ventricular hypertrophy and performance in hypertension. Am J Cardiol 1983,51:171-176
26. Narayan S, Devereux RB, Lutas EM, et a). Supernormal left ventricular
function in essential hypertension' evidence for increased contractility
[Abstract], J Am Coll Cardiol 1984;3:515
27. Henry WL, Clark CE, Epstein SE. Asymmetric septal hypertrophy:
echocardiographic identification of the pathognomonic anatomic abnormality of IHSS. Circulation 1973;47:225-233
II-17
28. Sahn DJ, DeMaria A, Kisslo J, Weyman A, the Committee on M-mode
Standardization of the American Society of Echocardiography. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic methods. Circulation 1978;58:
1072-1083
29. Henry WL, Clark CE, Epstein SE. Asymmetric septal hypertrophy
(ASH): the unifying link in the IHSS disease spectrum; observations
regarding its pathogenesis, pathophysiology and course. Circulation
1973;47:827-832
30. Morganroth J. Maron BJ, Henry WL, Epstein SE. Comparative left
ventricular dimensions in trained athletes. Ann Intern Med 1975;82:
521-524
31. Savage DD, Drayer J, Henry WL, etal. Echocardiographic assessment
of cardiac anatomy and function in hypertensive subjects. Circulation
19793:623-632
32. Henry WL, Ware J, Gardin JM, Hepner SI, McKay J, Weiner M.
Echocardiographic measurements in normal subjects. Circulation 1978;
57:278-285
33 Gardin JM, Henry WL, Savage DD, Ware JH, Bum C, Borer JS.
Echocardiographic measurement in normal subjects' evaluation of an
adult population without clinically apparent heart disease. JCU 1979;7:
439-447
34. Crawford MH, Grant D, O'Rourke RA, Starling MR, Groves BM.
Accuracy and reproducibility of new M-mode echocardiographic recommendations for measuring left ventricular dimensions. Circulation
1980;61:137-143
35. WoythalerTN, Singer SL, KwanOL, etal. Accuracy of echocardiography versus electrocardiography in detecting left ventricular hypertrophy: comparison with post mortem mass measurement. J Am Coll
Cardiol 1983;2:305-311
36 Devereux RB, Alonso DR, Lutas EM, Pickering TG, Harshfield GA,
Laragh JH Sensitivity of echocardiography for detection of left ventricular hypertrophy. In: ter Keurs HEDJ, Schipperheyn JJ, eds. Cardiac
left ventricular hypertrophy. The Hague: Martinus-Nijhoff, 1983;16—37
37. Gaash WH, Amonas WC, Levine HJ. Chronic aorticregurgitation:the
effect of aortic valvereplacementon left ventricular volume, mass and
function. Circulation 1978^8:825-836
38. Schiller NB, Skioldebrand CG, Schiller EJ, et al. Canine left ventricular
mass estimation by two-dimensional echocardiography. Circulation
1983;68:210-216
39. Helak JW, Reichek N. Quantitation of human left ventricular mass and
volume by two dimensional echocardiography. in vitro anatomic validation. Circulation 1981;63.1398-1407
40. Reichek N, Helak J, Plappert T, St John Sutton M, Weber KT. Anatomic validation of left ventricular mass estimates from clinical two-dimensional echocardiography: initial results. Circulation 1983;67:348-352
41. Vignola PA, Bloch A, Kaplan AD, Walker HJ, Chiotellis PN, Myers
GS Interobserver variability in echocardiography. JCU 1977^:238242
42 Monoson PA, O'Rourke RA, Crawford MH, White DH Measurements
of left ventricular wall thickness and systolic thickening by M-mode
echocardiography: interobserver and intrapatient variability. JCU 1978,
6:252-258
43. Pollick C, Fitzgerald PJ, Popp RL. Variability of digitalized echocardiography: size, source and means ofreduction.Am J Cardiol 1983^51.
576-582
44. Lapido GOA, Dunn FG, Pnngle TH, Bastian B, Lawrie TDV. Serial
measurement of left ventricular dimensions by echocardiography: assessment of week to week, inter- and intraobserver variability in normal
subjects and patients with valvular heart disease. Br Heart J 1980;44.
284-289
45. PietroDA, VoelkelGA, RayBJ, Parisi AF. Reproducibility of echocardiography: a study evaluating the variability of serial echocardiographic
measurements. Chest 1981;79:29-32
46. Devereux RB, Savage DD, Drayer JIM, Laragh JH. Left ventricular
hypertrophy and function in high, normal, and low-renin forms of essential hypertension. Hypertension 1982;4:524—531
47. Szlachcic J, Massie B, Greenberg B, Thomas D, Cheitlin M, Bristow
JD. Intertest variability of echocardiographic and chest x-ray measurements: implications for decision making in patients with aortic regurgitation. J Am Coll Cardiol 1986;7:1310-1317
48. UmaliMSA.PrakashR. Magnitude of daily variations in serial echocardiogupliii ucasureineiits of septalthickness, leftventncularwallthickness and estimated left ventricular mass [Abstract], Clin Res 1984;
32:212A
49. Panidis IP, Ross J, Ren JF, Kotler MN, Mintz GS. Comparison of
independent and derived M-mode echocardiographic measurement. Am
J Cardiol 1984;54:694-695
50. Henry WL, DeMaria A, Gramiak R, et al. Report of the American
Society of Echocardiography Committee on Nomenclature and Standards in Two-Dimensional Echocardiography. Circulation 1980;62:
212-217
II-18
ECHOCARDIOGRAPHY
51. Erbel R, Schweizer P, Lambertz H, et al. Echoventriculography: a
simultaneous analysis of two-dimensional echocardiography and cineventriculography. Circulation 1983;67:205-215
52. Henry WL, Griffith JM, Michaelis LL, Mclntosh CL, Morrow AG,
Epstein SE. Measurement of mitral orifice area in patients with mitral
valve disease by real time two-dimensional echocardiography. Circulation 1975;51:827-831
53. Nichol PM, Gilbert BW, Kisslo JA. Two-dimensional echocardiographic assessment of mitral stenosis. Circulation 1977;55:120—128
54. Lutas EM, Devereux RB, Borer JS, Goldstein JE. Echocardiographic
evaluation of mitral stenosis: a critical appraisal of its clinical value in
detection of severe stenosis and of valvular calcification. J Cardiovasc
Ultrasonogr 1983;2:131-139
55. Glover MV, Warren SE, Vieweg WVR, Ceretto WJ, Samtoy LM,
Hagan AO. M-mode and two-dimensional echocardiographic correlation with findings at catheterization and surgery in patients with mitral
stenosis. Am Heart J 1983;105:98-102
56. Motro M, Schneeweiss A, Lehrer E, Shmuel R, Neufeld HN. Correlation between cardiac catheterization and echocardiography in assessing
the severity of mitral stenosis. Int J Cardiol 1981; 1:25—34
57. Ditchey RV, Schuler G, Peterson KL. Reliability of echocardiographic
and electrocardiographic parameters in assessing serial changes in left
ventricular mass. Am J Med 1981;70:1042-1050
SUPPL II HYPERTENSION, VOL 9, No 2, FEBRUARY
1987
58. Wyatt HL, Haendchen RV, Meerbaum S, Corday E. Assessment of
quantitative methods for two-dimensional echocardiography. Am J Cardiol 1983;52:396-401
59. Skorton DJ, McNary CA, Child JS, Newton FC, Shah PM. Digital
imaging processing of two-dimensional echocardiograms: identification
of the endocardium. Am J Cardiol 1981;48:479-486
60. Kantrowitz NE, Schnittger I, Schwarzkopf A, Fitzgerald PJ, Popp RL.
Rapid, semi-automated technique for estimating left ventricular volume. Am Heart J 1983;106:521-527
61. Gordon EP, Schnittger I, Fitzgerald PJ, Williams P, Popp RL. Reproducibility of left ventricular volumes by two-dimensional echocardiography. J Am Coll Cardiol 1983;2:506-513
62. Wallerson DC, Devereux RB. Reproducibility of quantitative echocardiography: factors affecting variability of imaging and Doppler measurements. Echocardiography: a Review of Cardiovascular Ultrasound
1986;3:219-235
63. Unverferth DV, Lewis RP, Leier CV, Magorien RD, Fulkerson PK.
Reproducibility of echocardiography in chronic congestive heart failure. JCU 1980;8:479-482
64. MacMahon SW, Hickey AT, Wilcken DEL. M-mode echocardiography: reproducibility of serial left ventricular measurements in subjects
with normal ventricles and patients with congestive cardiomyopathy.
Aust NZ J Med 1983;13:457-462
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
Reproducibility of echocardiographic left ventricular measurements.
D C Wallerson and R B Devereux
Downloaded from http://hyper.ahajournals.org/ by guest on May 7, 2017
Hypertension. 1987;9:II6
doi: 10.1161/01.HYP.9.2_Pt_2.II6
Hypertension is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1987 American Heart Association, Inc. All rights reserved.
Print ISSN: 0194-911X. Online ISSN: 1524-4563
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://hyper.ahajournals.org/content/9/2_Pt_2/II6
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Hypertension 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 Hypertension is online at:
http://hyper.ahajournals.org//subscriptions/