From Noninvasive Cardiac Imaging Laboratories, Section of Cardiology,
Department of Medicine, The University of Chicago Medical Center, Chicago,
Ill.
Correspondence to Roberto M. Lang, MD, MC5084, University of Chicago Medical Center, 5841 S Maryland Ave, Chicago, IL 60637. E-mail rlang{at}medicine.bsd.uchicago.edu
Methods and Results—Color Kinesis images and mitral and
pulmonary vein flow Doppler data were acquired in 29
patients with LV hypertrophy and 29 age-matched control
subjects. In addition, Color Kinesis data were correlated to
coronary angiographic findings in 15 patients with suspected
coronary artery disease. Segmental analysis of Color
Kinesis images was used to obtain time histograms of regional
diastolic fractional area change, wherein early and late
peaks (peaks 1 and 2) reflected rapid LV filling and atrial
contraction, respectively. Regional mean LV filling time and filling
curves were used to objectively identify diastolic
endocardial motion asynchrony in patients with LV
hypertrophy and coronary artery disease. None of
the mitral and pulmonary vein Doppler indices
differentiated patients with normalized mitral Doppler profile
(n=13) from control subjects, whereas reduced peak1/peak2 ratio and
prolonged mean filling time indicated augmented contribution of atrial
contraction toward LV filling (P<.05). In 22 of 25
patients with LV hypertrophy and preserved systolic
function and in all patients with coronary artery disease,
delayed diastolic endocardial motion was observed in at
least one segment.
Conclusions—Analysis of Color Kinesis images provides
objective assessment of global and regional LV filling properties and
allows identification of both diastolic dysfunction in
patients with normalized Doppler indices and wall motion
asynchrony.
Moreover, Doppler flow profiles provide information on global
rather than regional LV diastolic properties. The ability
to assess regional diastolic abnormalities, which may
affect overall ventricular filling, depends on
diagnostic techniques that require injection of contrast or
radiopharmaceutical substances.4 5 6 Color Kinesis
(Hewlett Packard) is an echocardiographic technique
that allows objective evaluation of regional systolic wall
motion.7 The aim of this study was to determine
the feasibility of the use of Color Kinesis to assess global as well as
regional LV diastolic properties. Accordingly,
diastolic endocardial motion data obtained from Color
Kinesis images acquired in patients with LV hypertrophy and
age-matched control subjects were compared with traditional Doppler
indices. In addition, the ability of Color Kinesis to objectively
characterize regional LV diastolic asynchrony was evaluated
in patients with LV hypertrophy and in patients undergoing
coronary angiography for suspected coronary artery
disease.
Study Population
Data Acquisition
Doppler Echocardiography
Color Kinesis Data Acquisition
Data Analyses
Doppler Analysis
Analysis of Color Kinesis Images
Study Design
Protocol 2
Protocol 3
Statistical Analysis
Fig 2
Fig 3
Characterization of LV Filling Profiles
When compared with control subjects, patients in groups 1 and 3
exhibited distinct diastolic endocardial motion patterns
(Fig 4
Time histograms obtained in group 2 exhibited a disproportionately
large contribution of atrial contraction toward LV filling despite
normalized mitral Doppler profiles (normal E/A ratios with
prolonged deceleration time, n=7; normal E/A ratios with normal
deceleration time, n=6) (Fig 4
Protocol 2
The patterns of regional LV endocardial motion were consistent
and relatively uniform in normal subjects, as demonstrated by the
homogeneous profiles of regional time curves (Fig 6
Protocol 3
Our study was designed to determine whether regional quantitative
analysis of Color Kinesis images7 10
could provide (1) information additional to Doppler indices of
global LV diastolic function, particularly in patients with
concentric hypertrophy and normalized mitral inflow
profiles and (2) objective characterization of regional
diastolic LV wall motion.
Quantitative Analysis of End-diastolic Color
Kinesis Images
To evaluate the temporal patterns of diastolic endocardial
motion on a regional basis, segmental fractional area change was also
displayed as time curves. The percentage of total endocardial motion
completed at 50% of the LV filling period was computed for each
segment and used to objectively assess regional diastolic
asynchrony during rapid LV filling. Delayed diastolic
endocardial motion was identified by comparing individual curves
obtained in hypertensive patients with reference profiles obtained in
normal subjects (Figs 7
Assessment of Global LV Diastolic Function
In contrast, the analysis of Color Kinesis images, which
provides information on LV diastolic properties by
quantifying endocardial motion,26 27 28 separated
patients with LV hypertrophy from control subjects.
Although the magnitude of diastolic fractional area change
was not significantly different between hypertensive and normal
subjects, LV filling was delayed in patients with concentric
hypertrophy, as evidenced by the increased endocardial
motion during late diastole (Fig 3
Characterization of LV Filling Profiles
In these patients, Color Kinesis time histograms depicted a decreased
early peak, whereas the late peak of diastolic fractional
area change was augmented, resulting in diminished peak1/peak2 ratios
with prolonged mean LV filling times (Table 3
Color Kinesis time histograms were abnormal in group 2 patients,
whereas Doppler indices were similar to those obtained in
age-matched control subjects. In addition, mean filling time obtained
from Color Kinesis images was prolonged in these patients (Table 3
Assessment of Regional Diastolic Function
Regional diastolic LV filling abnormalities are
sensitive early signs of myocardial ischemia and may occur when
systolic function is still
preserved.5 34 35 Thus characterization of
regional LV diastolic endocardial motion with the use of
Color Kinesis appears to be of potential value for the detection of
myocardial ischemia. In this study, patients with
nonsignificant coronary artery stenosis had an index of
asynchrony similar to that observed in normal subjects. In contrast,
patients with coronary artery disease exhibited significantly
increased regional diastolic wall motion
heterogeneity. As opposed to patients with LV
hypertrophy in whom segmental diastolic
abnormalities were identified without predominance of specific regions,
in patients with coronary artery disease, delayed endocardial
diastolic motion was specific to myocardial segments
perfused by narrowed epicardial coronary arteries. Absence of
delayed endocardial motion in certain LV segments supplied by narrowed
epicardial coronary arteries may be explained by the absence of
significant myocardial ischemia at rest in these myocardial
regions. Further studies are needed to determine whether
characterization of regional diastolic endocardial motion
with the use of Color Kinesis may help to noninvasively assess
myocardial ischemia during stress testing.
Several diagnostic methods such as
contrast4 or radionuclide
ventriculography5 6 35 have been used to quantify
regional diastolic function by generating regional
diastolic volume–time curves. Color Kinesis has several
advantages over these modalities. First, Color Kinesis does not require
exposure to radiation and can therefore be used readily for serial
assessment of regional diastolic function. Second, with the
use of Color Kinesis, it is possible to quantify regional
diastolic endocardial motion on line.
Limitations
Conclusions
Received July 9, 1997;
revision received October 20, 1997;
accepted December 1, 1997.
© 1998 American Heart Association, Inc.
Clinical Investigation and Reports
Quantitative Evaluation of Global and Regional Left Ventricular Diastolic Function With Color Kinesis
Abstract
Top
Abstract
Introduction
Methods
Results
Discussion
References
Background—Diastolic
wall motion asynchrony is a major determinant of impaired left
ventricular (LV) filling in patients with concentric
hypertrophy and coronary artery disease. We
evaluated the ability of Color Kinesis, a new
echocardiographic technique that color-encodes
endocardial motion, to quantitatively assess global and regional LV
filling properties.
Key Words: echocardiography • diastole • hypertrophy • coronary disease
Introduction
Top
Abstract
Introduction
Methods
Results
Discussion
References
Abnormalities in left
ventricular (LV) diastolic filling often
precede systolic dysfunction in various disease states such as
cardiac hypertrophy1 and
coronary artery disease.2 LV
diastolic properties are evaluated indirectly by
measurement of transmitral flow velocities with pulsed Doppler
echocardiography. Distinct mitral inflow patterns
have been used to differentiate abnormal relaxation from restricted
filling.3 However, the diagnosis of
diastolic dysfunction in patients with normalized mitral
inflow patterns3 remains challenging.
Methods
Top
Abstract
Introduction
Methods
Results
Discussion
References
Color Kinesis: Principles of Operation
Color Kinesis analyzes regional backscatter in each
acoustic frame in real time and classifies each pixel as either blood
or myocardial tissue. Pixel transitions from tissue to blood during
diastole are detected and color-encoded (Fig 1
). Color overlays are updated on a
frame-by-frame basis by adding one color at a time (30 frames/s). Thus
a single end-diastolic frame provides an integrated display
of the timing and magnitude of endocardial wall motion.
View larger version (41K):
[in a new window]
Figure 1. Principles of operation of Color Kinesis (A and
B). Segmentation schemes used in the short-axis and apical four-chamber
views (C). Ant indicates anterior; asp, anteroseptal; sp, septal; inf,
inferior; pst, posterior; lat, lateral;
b-lat, basal lateral; m-lat, mid-lateral; a-lat, apical
lateral; a-sp, apical septal; m-sp, mid-septal; and b-sp, basal
septal.
Out of 100 screened subjects, 85 with good
echocardiographic image quality and adequate tracking
by Color Kinesis of >80% of LV endocardial boundary were selected for
the study. From these, 12 patients were excluded because of (1)
pericardial effusion and previous pericardotomy (n=3); (2) all types of
arrhythmias and conduction abnormalities (n=2); (3) heart rates
<55 or >100 bpm (n=2); (4) moderate to severe mitral or aortic
regurgitation assessed with color-flow Doppler
(n=2); (5) mitral or aortic valve stenosis (n=2); and (6)
dynamic LV outflow tract obstruction (n=1). Finally, 29 patients (21
men, 8 women; age, 55±14 years; range, 30 to 81) with concentric LV
hypertrophy secondary to long-standing systemic
hypertension were recruited into the study as well as 29 age-matched
normal subjects (13 men, 16 women; age, 54±14 years; range, 32 to 88)
who were used as control subjects. LV hypertrophy was
defined as LV mass index >2 SD above age- and sex-normalized
values.8 Concentric LV hypertrophy
was defined as septal thickness <1.3 times the thickness of the
posterior wall when measured with M-mode in the parasternal short-axis
view. Control subjects had (1) no history of
cardiovascular disease, (2) normal two-dimensional
echocardiographic study, and (3) LV mass index and
mitral flow velocity pattern (E/A ratio and deceleration time) within
the normal range.8 9 In addition, 15 patients who
underwent coronary angiography for suspected coronary
artery disease (10 men, 5 women; age, 61±13 years; range, 45 to 84)
were studied. Significant coronary stenosis defined as
a reduction 
50% of the arterial lumen was found in 7
patients (1 single-vessel, 3 double-vessel, and 3 triple-vessel
disease). The remaining 8 patients had nonsignificant coronary
artery atherosclerosis.
In all study subjects, a complete transthoracic
echocardiographic study including M-mode,
two-dimensional imaging, pulsed-wave Doppler, and color flow
mapping of valvular orifices was performed with either a 2.5-
or 3.5-MHz transducer (SONOS 2500, Hewlett-Packard).
In all subjects, two-dimensional parasternal short-axis views
were obtained at the mid papillary muscle level, followed by apical
four-chamber views, and recorded on videotape for off-line
measurements of LV mass index. Two-dimensional targeted M-mode imaging
of the left ventricle was performed in the short-axis view to measure
end-diastolic wall thickness. Mitral and pulmonary
vein flow pulsed-wave Doppler velocity profiles were acquired
during passive end-expiration, as previously
described.9
Diastolic Color Kinesis images were obtained
from the LV mid papillary short-axis and apical four-chamber views as
described previously.10 The timing of color
encoding was set to begin at the first frame in which outward
endocardial motion was noted, and its duration was set to its maximal
value (19 frames). To minimize the effects of beat-to-beat variability,
two nonconsecutive end-diastolic Color Kinesis images were
acquired in each view and stored on an optical disk.
Two-dimensional and M-Mode Measurements
LV mass index was measured with the two-dimensional area-length
method.11 Long-axis LV length was measured from
the apical four-chamber view as the distance between the mid mitral
annulus plane and the tip of the apical endocardial border. Both LV
chamber dimensions and wall thickness were measured with the
conventional leading edge technique.
Measurements were performed off-line and values were
obtained as the mean of three nonconsecutive beats. Isovolumic
relaxation time was measured as the time interval between the aortic
valve closure and the onset of the mitral valve inflow. Peak velocities
during rapid filling (E) and atrial contraction (A) as well as the area
under each peak (VTIE and
VTIA) were measured. Subsequently, the early to
late diastolic mitral flow velocity E/A ratio and
VTIE/VTIA ratio were
calculated, and the deceleration time was measured. Only cardiac cycles
with linear deceleration and clearly defined peaks were
used.3 In addition, we measured the time to half
LV filling corresponding to 50% of the area under the mitral velocity
profile. Pulmonary vein inflow velocities were measured at
end-expiration. The ratio between peak forward flow velocity during
ventricular systole and diastole (S/D ratio) as
well as peak retrograde velocity at atrial contraction (A-pv) were
measured.
End-diastolic color-encoded images were divided into
six segments (Fig 1C
) with previously described custom
software.7 10 In each segment, colored pixel
counts were used to calculate regional fractional area change, which
was displayed as stacked time histograms. This display allowed clear
identification of peak area change during rapid LV filling (peak 1) and
during atrial contraction (peak 2). To facilitate intersubject
comparisons, we used linear interpolation to obtain 20 values of
segmental fractional area change in 5% increments of LV filling time.
Thus irrespective of heart rate, LV filling time was 100% in each
subject. The peak1/peak2 ratio was measured, and mean time of LV
filling was computed for each segment and displayed as a bar
diagram.10 Fractional area change was integrated
with respect to time and normalized to 100% in each segment. This
normalization eliminated intersegmental differences in magnitude of
regional endocardial motion, which was displayed as time curves
reaching 100% of endocardial motion at 100% filling time. For each
subject, data obtained from two Color Kinesis images in each view were
averaged. Composite regional time curves obtained in all normal
subjects were used as reference patterns of regional LV filling.
Protocol 1
To determine the clinical value of Color Kinesis for the
evaluation of global LV diastolic function, Doppler and
Color Kinesis data were initially compared between patients with
concentric LV hypertrophy and age-matched control subjects.
Subsequently, patients were divided into three subgroups on the basis
of their Doppler mitral inflow patterns3 by
comparison of individual E/A ratios and deceleration times with the
previously reported 95% confidence intervals, obtained in a normal
population over a wide range of ages.9 Group 1
included 12 patients with an abnormal relaxation pattern (reduced E/A
ratios and prolonged deceleration times). Group 2 included 13 patients
with age- and sex-normalized E/A ratio. Group 3 included 4 patients
with a restrictive filling pattern (increased E/A ratios with short
deceleration times). In each group, Color Kinesis data were compared
with conventional Doppler-derived indices. In addition, Color
Kinesis and Doppler data were compared between groups and
corresponding control subjects.
To determine the ability of Color Kinesis to objectively assess
regional diastolic function, the proportion of
diastolic endocardial motion completed after the first half
of LV filling time was determined for each regional time curve. These
values were averaged and standard deviations used as an index of
diastolic asynchrony. Data were compared between patients
with LV hypertrophy and preserved systolic function
(n=25) and control subjects. Regional curves obtained in normal
subjects were averaged and a reference pattern of endocardial motion
was defined in each segment as 1 SD around the mean. Delayed segmental
endocardial motion was defined as decreased percentage of
motion completed at 50% of filling time when compared with normal
values.
To evaluate the ability of Color Kinesis to detect delayed
diastolic endocardial motion in LV regions supplied by
coronary arteries with various degrees of stenosis,
regional motion was correlated with perfusion territories on a
segmental basis. As in protocol 2, the index of diastolic
wall motion asynchrony was calculated from Color Kinesis images.
Regional delayed endocardial motion was defined as a downward shift of
individual segmental Color Kinesis time curves relative to the
corresponding reference profile at 50% of the total filling time. The
presence or absence of delayed endocardial motion was correlated with
coronary angiographic findings on a segmental
basis.11
Intergroup comparisons were performed with either the
Student's t test or the Mann-Whitney rank-sum test whenever
data failed the normality test. Nonparametric Spearman
correlation coefficients were computed and tested for differences
between the Doppler
VTIE/VTIA ratios and Color
Kinesis peak1/peak2 ratios (P
.05 considered
significant).
Results
Top
Abstract
Introduction
Methods
Results
Discussion
References
Protocol 1
Assessment of Global LV Diastolic Function
LV mass index was greater in patients with hypertrophy
when compared with normal subjects (178±52 versus 67±14
g/m2, P<.0001). No intergroup
differences were noted in heart rate (71±12 versus 65±9 bpm), mitral
inflow E/A ratio, time to half LV filling, pulmonary venous
flow S/D ratio, and A-pv (Table 1). In
contrast, isovolumic relaxation and deceleration times were prolonged,
and VTIE/VTIA ratio
significantly decreased in patients with LV hypertrophy
(Table 1).
View this table:
[in a new window]
Table 1. Pulsed-Wave Doppler-Derived Indices and
Parameters of Global Diastolic Function
Obtained From Color Kinesis Data Obtained in Patients With Concentric
Left Ventricular Hypertrophy and Age-Matched
Control Subjects shows an example of
end-diastolic Color Kinesis images obtained in a patient
with concentric LV hypertrophy and an age-matched control
subject. The magnitude of diastolic endocardial motion was
similar between both subjects. However, despite similar heart rates,
the patient exhibited more late colors (yellow and orange), reflecting
augmented contribution of atrial contraction toward LV filling. Even
though mean regional diastolic fractional area change was
not significantly different between patients with LV
hypertrophy and control subjects (74±9% versus 68±19%
and 46±10% versus 47±15% in the short-axis and apical four-chamber
views, respectively), the proportion of late diastolic
filling was greater in patients with LV hypertrophy.
View larger version (136K):
[in a new window]
Figure 2. End-diastolic Color Kinesis images
obtained in a patient with concentric left ventricular
hypertrophy (LVH) and an age-matched normal subject (CTL),
in the short-axis (top) and apical four-chamber (bottom) views. Note
the presence of thick yellow and orange bands in the patient with LVH
in late diastole (see color scale in Fig 1 ), which reflects
the greater dependence of left ventricular filling on
atrial contraction. In contrast, in the normal subject, blue and green
colors corresponding to early diastolic filling are
predominant. shows time histograms of
regional diastolic fractional area change averaged for
patients with LV hypertrophy and for age-matched control
subjects. The ratio of the early to late peaks (peak1/peak2) was lower
in patients (Table 1), while the timing of both peaks was similar
between groups. The mean LV filling time was prolonged in patients with
LV hypertrophy.
View larger version (65K):
[in a new window]
Figure 3. . Averaged time histograms depicting regional
fractional area change (RFAC) in percentage of global
end-diastolic area (EDA) as a function of filling time in
patients with concentric left ventricular
hypertrophy (LVH) (right) and age-matched control subjects
(CTL, left). Each layer depicts the contribution of one segment. The
dashed band represents 1 SD of the mean. In both the short-axis
and apical four-chamber views, patients with LVH exhibited a larger
contribution of atrial contraction compared with CTL, as reflected by a
higher peak in late diastole. Also see Fig 1
abbreviations.
With the exception of patients with a restrictive pattern (group
3), all patients and control subjects had normal LV systolic
function. Mitral and pulmonary venous flow Doppler indices
obtained in the three subgroups of patients and corresponding control
subjects are shown in Table 2. The
isovolumic relaxation time was prolonged in patients with abnormal
relaxation (group 1) and normalized mitral flow Doppler (group 2)
and decreased in patients with a restrictive pattern (group 3). The
time to half LV filling was increased in group 1 and decreased in group
3, whereas it was similar to control subjects in group 2. The
pulmonary vein S/D ratio was increased in group 1 and decreased
in group 3, relative to control subjects. The A-pv was only
nonsignificantly augmented in group 3. In contrast, in patients with
normalized mitral inflow profiles, none of these Doppler
parameters were statistically different from control
subjects (Table 2).
View this table:
[in a new window]
Table 2. Pulsed-Wave Doppler-Derived Indices Obtained in
Three Subsets of Patients With Concentric Left Ventricular
Filling Hypertrophy and in Corresponding Age-Matched
Control Subjects ). Patients with abnormal
relaxation exhibited augmented atrial contribution toward LV filling.
In contrast, insignificant atrial contribution was noted in patients
with restrictive mitral inflow profiles. The time histograms were
similar to mitral inflow profiles in normal subjects and in groups 1
and 3 (Fig 4). Peak1/peak2 ratio was diminished in patients with
abnormal relaxation (Fig 5, left) and
markedly increased in patients with restrictive patterns (Fig 5, right). This ratio correlated with
VTIE/VTIA ratio in both
views (r=.81 and .73, respectively; P<.001).
Mean LV filling time was prolonged in group 1 and shortened in group 3
(Table 3).
View larger version (55K):
[in a new window]
Figure 4. Examples of end-diastolic Color
Kinesis images (top) with the time histograms (bottom) obtained in four
subjects with distinct mitral Doppler flow profiles (middle). When
compared with the normal subject (left), the patient with abnormal
relaxation (second column) exhibited predominant late
diastolic colors, resulting in an inverted peak1/peak2
ratio, which reflected a markedly increased contribution of atrial
contraction toward left ventricular filling. The patient
with normalized Doppler profile (third column) exhibited similar
Color Kinesis pattern. In the patient with restrictive filling pattern
(right), the presence of blue colors indicated predominant early left
ventricular filling with negligible atrial contribution.
Also see Figs 1 and 3 abbreviations.
View larger version (64K):
[in a new window]
Figure 5. Same data as in Fig 3 shown for three
subgroups of patients. Increased fractional area change during late
diastole, consistent with augmented atrial
contribution, was evident in both the short-axis and apical
four-chamber views in group 1 (abnormal relaxation) and group 2
(normalized mitral inflow velocity profile). Group 3 patients
(restriction) exhibited predominant early diastolic left
ventricular filling. Also see Figs 1 and 3
abbreviations.
View this table:
[in a new window]
Table 3. Averaged Color Kinesis Parameters of
Left Ventricular Diastolic Function Obtained in
Three Subsets of Patients With Cardiac Hypertrophy and in
Corresponding Age-Matched Normal Control Subjects ) and normal pulmonary vein
Doppler parameters (Table 2). The atrial contribution
in this group was similar to group 1 (Fig 5, middle). Accordingly,
peak1/peak2 ratios were decreased and mean LV filling time prolonged
(Table 3). As a result, in these patients,
VTIE/VTIA ratio correlated
poorly with the Color Kinesis peak1/peak2 ratio (r=.43;
P=.02).
In patients with LV hypertrophy and preserved
systolic function (n=25), the proportion of
diastolic fractional area change completed during the first
half of the LV filling in both the short-axis and apical four-chamber
views was reduced when compared with control subjects (62±12% versus
74±7% and 66±9% versus 76±5%, respectively, P<.001).
Regional diastolic endocardial motion was more
heterogeneous in patients with LV hypertrophy,
as reflected by increased index of asynchrony (10.4% versus 6.9%,
P<.01, short-axis; 11.2% versus 8.9%, P=.05,
apical four-chamber views). , top). In contrast, patients with LV
hypertrophy frequently exhibited diastolic
endocardial motion asynchrony (Fig 6, middle), corroborated by the wide
intersegmental variability in the regional LV filling times (Fig 6, bottom). Although groups 1 and 2 had normal LV systolic
performance, 22 of 25 (88%) patients exhibited delayed
endocardial motion during early diastole (Fig 7) in at least one segment (97 of 300
total segments). Late diastolic endocardial motion occurred
most frequently in the septal, anteroseptal, and anterior segments in
the short-axis view (13, 14, and 11 segments, respectively) and in the
basal septal, mid septal, and lateral segments in the apical
four-chamber view (7, 8, and 19 segments, respectively).
View larger version (34K):
[in a new window]
Figure 6. Assessment of diastolic endocardial
motion asynchrony with Color Kinesis. In a normal subject (CTL, top),
regional left ventricular (LV) filling time curves were
relatively uniform in both the short-axis (SAX) and apical four-chamber
(A4C) views. In contrast, a patient with concentric
hypertrophy exhibited diastolic endocardial
motion asynchrony, reflected by inhomogeneities of regional filling
curves (LV hypertrophy, LVH; middle) and regional mean
filling times (<t>, bottom). Bar histograms show the individual
regional mean LV filling time obtained in this patient (open bars)
compared with the average values obtained in 29 normal subjects (solid
bars with SD graphed). Note that in specific segments, mean LV filling
time exceeds 1 SD around the mean obtained in the normal subjects. Also
see Figs 1 and 3 abbreviations.
View larger version (32K):
[in a new window]
Figure 7. Example of regional filling curves obtained in a
patient with concentric left ventricular
hypertrophy (LVH) and normal systolic function. For
each segment, a reference profile of diastolic endocardial
motion was generated by averaging curves obtained in normal subjects
(n=29); dashed lines represent 1 SD around the mean. In this
case, the curves in the anterior, anteroseptal, and septal segments
reflect delayed endocardial motion during early diastole.
For each segment, mean time of filling (in milliseconds) is displayed
(open bars) together with the normal values (solid bars) and SD for
comparison. Also see Fig 3 abbreviations.
The index of asynchrony obtained in patients with nonsignificant
coronary artery lesions was not significantly different from
that of the control group (9.4% versus 6.9%, short-axis; and 8.5%
versus 8.9%, apical four-chamber views; NS). In contrast, in patients
with coronary artery disease this index was significantly
increased (17.1% versus 6.9%, short-axis; and 12.9% versus 8.9%,
apical four-chamber views; P<.05) when compared with the
control group because of delayed regional endocardial motion.
Similarly, patients with coronary artery disease had larger
index of asynchrony compared with patients with nonsignificant
coronary artery lesions (17.1% versus 9.4%, short-axis; and
12.9% versus 8.5%, apical four-chamber views; P.05). In
patients with coronary artery disease, diastolic
endocardial motion was delayed in 28 of 84 segments studied (33%) (Fig 8), whereas in patients with
nonsignificant coronary lesions, regional diastolic
abnormalities were only observed in 8 of 96 segments analyzed
(8%). Coronary angiograms revealed that 70 LV segments were
supplied by narrowed coronary epicardial arteries (50% of
arterial lumen). Delayed regional diastolic
endocardial motion was observed in 27 of these segments and remained
within the normal profile in 43 LV segments. A total of 110 LV segments
were perfused by coronary arteries with nonsignificant lesions.
In 9 of these segments, Color Kinesis analysis detected delayed
endocardial motion, whereas in 101 LV segments, normal regional
diastolic displacement was noted.
View larger version (33K):
[in a new window]
Figure 8. Example of regional filling curves obtained in the
short-axis view in a patient with a 75% stenosis of the mid
left anterior descending coronary artery. When compared with
the normal profiles (dashed bands), diastolic endocardial
motion is substantially delayed in the anteroseptal and septal left
ventricular segments. Also see Figs 1 and 3
abbreviations.
Discussion
Top
Abstract
Introduction
Methods
Results
Discussion
References
Considerable efforts have been recently directed toward the
development of noninvasive techniques capable of accurate and serial
assessment of LV diastolic properties. However, currently
available methods such as Doppler and acoustic
quantification12 13 only provide information on
global rather than regional LV filling. Moreover, diastolic
dysfunction may be difficult to identify in patients with normalized
Doppler indices.14 Global LV
diastolic function is adversely affected by heterogeneities
in regional contraction and relaxation.15 16
Consequently, regional diastolic properties need to be
fully elucidated to better understand this relationship.
To compare mitral flow Doppler-derived indices with Color
Kinesis data, we limited the analysis of the latter to the
period of LV filling. Color Kinesis images provide real-time
information on the timing and magnitude of diastolic
endocardial displacement. Using segmental analysis of these
images,7 10 we calculated various
parameters of regional diastolic endocardial
motion. The segmentation schemes used (Fig 1) were similar to those
recommended by the American Society of
Echocardiography.11 To
facilitate the comparison between LV inflow Doppler and Color
Kinesis findings, regional diastolic fractional area change
was displayed as a function of time. Interpolation allowed intersubject
comparisons of diastolic endocardial motion profiles
irrespective of the duration of LV filling. These time histograms had
two distinct peaks, one in early diastole corresponding to
the period of rapid LV filling and a second one in late
diastole reflecting the atrial contribution toward LV
filling (Figs 3, 4, and 5). and 8).
As expected, both the isovolumic relaxation and early deceleration
times were prolonged and the
VTIE/VTIA ratio reduced in
patients with LV hypertrophy when compared with age-matched
control subjects (Table 1), reflecting impaired LV relaxation and
compliance.17 However, normal subjects and
patients with LV hypertrophy had similar E/A ratios, time
to half LV filling, pulmonary vein S/D ratios, and peak
reversal flow velocities during atrial contraction (Table 1). The
latter findings differ from those reported in previous studies in which
reduced E/A ratios18 and augmented S/D
ratios19 20 were reported in patients with
cardiac hypertrophy as a result of increased dependence on
atrial systolic filling. These discrepancies might reflect
differences in age, loading conditions, and heart rates, which are
known to affect both mitral and pulmonary vein flow velocity
profiles.9 21 22 23 24 In our study, however, no
differences in heart rate or age were present between patients with
LV hypertrophy and control subjects. Because it is now well
established that Doppler flow profiles change with the progression
of LV hypertrophy,3 20 25 differences
in both the severity and duration of hypertensive heart disease were
presumably responsible for the inability of the Doppler indices to
differentiate between hypertensive and normal subjects. ). This finding was
confirmed by a decreased peak1/peak2 ratio in patients with cardiac
hypertrophy, in agreement with decreased Doppler
VTIE/VTIA ratio (Table 1)
and with previous studies based on radionuclide
techniques26 27 and acoustic
quantification.12 In contrast, mean filling time
obtained from Color Kinesis was prolonged, whereas the time to half LV
filling derived from Doppler was similar in patients with
hypertrophy and control subjects (Table 1).
Progression of hypertensive heart disease modifies LV
diastolic properties, which in turn results in gradual
alterations of mitral flow Doppler
patterns,25 from abnormal relaxation (group 1) to
normalized (group 2) and finally restrictive (group 3) profiles.
Interestingly, the pulmonary vein flow S/D ratio as well as the
time to half LV filling progressively decreased from group 1 to group
3, whereas the E/A ratio exhibited opposite changes (Table 2). These
findings suggest that the proportion of LV filling that depends on
atrial contraction is limited in hypertensive patients with advanced LV
hypertrophy.29 As previously
reported,14 25 hypertensive patients with
normalized mitral flow velocity profiles and age-matched control
subjects had similar patterns of pulmonary venous flow
velocities (Table 2). Consequently, none of the conventionally used
Doppler indices identified global diastolic dysfunction
in our patients with LV hypertrophy and normalized mitral
flow velocity patterns. ). Similar findings were
noted in patients with abnormal relaxation (Figs 4 and 5), in agreement
with previous studies.12 27 In contrast, patients
with restrictive mitral flow Doppler patterns exhibited a markedly
increased peak1/peak2 ratio (Figs 4 and 5) with short mean filling
times (Table 3). Thus Color Kinesis time histograms depicted two
distinct abnormal patterns: (1) a reduced contribution of rapid LV
filling compensated by augmented atrial contraction in patients with
either abnormal relaxation or normalized mitral Doppler profiles
and (2) a predominant rapid filling and impaired left atrial emptying
observed in patients with restrictive LV inflow velocity profiles. ),
whereas the time to half LV filling failed to separate group 2 patients
from control subjects (Table 2). This discrepancy between instantaneous
variations in LV inflow velocities and diastolic fractional
area change can be explained as follows. First, using quantitative
evaluation of mitral flow propagation along the LV long axis with color
M-mode Doppler, Takatsuji et al30 have
recently demonstrated that LV filling flow propagation is rapidly
attenuated despite increased early transmitral velocities in patients
with normalized Doppler profiles. This finding was explained by the
marked increase in ventricular pressure immediately after
mitral valve opening secondary to reduced LV compliance, which might
stall the driving force of LV filling near the mitral
orifice.30 Second, it has been demonstrated that
mitral flow velocity does not have a flat spatial
profile.31 Third, blood flow entering the left
ventricle appears to be oriented differently in early versus late
diastole.22 Finally, the mitral
annulus cross-sectional area has been shown not only to gradually
increase as diastole
progresses22 32 33 but also to change in
shape.33
Nonuniformity in temporal and regional distribution of load
and inactivation has been described as a major factor influencing LV
relaxation.15 These local diastolic
disturbances can be identified as deviations of endocardial
motion from normal patterns, which are also nonuniform to a certain
degree.15 Accordingly, we used segmental
analysis of end-diastolic Color Kinesis images to
characterize diastolic wall motion asynchrony in patients
with concentric LV hypertrophy and in normal subjects for
comparison. In patients with cardiac hypertrophy, regional
endocardial motion occurred proportionally later in
diastole in all LV segments when compared with control
subjects. This was reflected by the smaller percentage of total
diastolic fractional area change completed within the first
half of the LV filling period. In normal subjects,
diastolic endocardial motion was relatively uniform in all
segments. In agreement with previous studies6 in
which increased diastolic wall motion nonuniformity in
patients with cardiac hypertrophy has been reported in
association with impaired global LV relaxation, our patients exhibited
diastolic wall motion asynchrony evidenced by delayed
endocardial motion in 32% of LV segments (Fig 7).
Diastolic wall motion asynchrony may be related to various
mechanisms, such as heterogeneous myocardial
hypertrophy resulting in regional differences in wall
stress, localized foci of interstitial fibrosis, and
nonuniform inactivation secondary to the loss of contractile elements
and normal intercellular connections.15
Consequently, increased diastolic wall motion nonuniformity
may result in uncoordinated and consequently prolonged LV filling in
patients with cardiac hypertrophy. Although in our study,
patients with diastolic wall motion asynchrony had normal
systolic function, nonuniformity in systolic wall
motion not detectable by conventional visual interpretation of
two-dimensional echocardiograms may have contributed to the regional
diastolic abnormalities.5 Bonow et
al6 have reported that only 15% of patients with
cardiac hypertrophy and preserved systolic function
exhibited homogeneous LV filling when evaluated with
radionuclide angiography. Similarly, in our study, only 3 of 25
patients (12%) with LV hypertrophy and preserved
systolic function had no evidence of diastolic
asynchrony.
Color Kinesis has several limitations, such as its
dependence on image quality and operator-dependent gain
settings.7 10 Color Kinesis does not allow color
encoding of the entire LV filling sequence in patients with heart rate
<55 bpm because the diastolic color scale currently has
only 19 hues. In addition, the low temporal resolution of Color Kinesis
(30 frames/s) may not allow accurate enough definition of endocardial
motion at high heart rates. For all these reasons, almost one third of
our screened patients were not suitable for adequate Color Kinesis
studies. Further studies are needed to determine the feasibility and
value of quantitative analysis of diastolic Color
Kinesis images in nonselected patients with LV hypertrophy
or ischemic heart disease. Some of these limitations could be
overcome by using higher frame-rate imaging in conjunction with an
extended color scale and automated gain settings. Because Color Kinesis
is designed to color-encode endocardial motion, it is not suited for
the assessment of the isovolumic relaxation period. In addition,
because coronary angiograms were not obtained in our patients
with LV hypertrophy, the possibility of underlying
coronary artery disease in these patients could not be ruled
out. Finally, in our study, patients with LV hypertrophy
and a restrictive filling pattern constituted a small group.
Accordingly, our results need to be confirmed in a larger
population.
In this study, quantitative analysis of Color Kinesis
images was used to (1) identify diastolic dysfunction in
patients with concentric cardiac hypertrophy, with
normalized mitral inflow Doppler profiles and (2) characterize and
quantify diastolic wall motion asynchrony in patients with
cardiac hypertrophy and in patients with coronary
artery disease. Our results demonstrate that quantitative
analysis of Color Kinesis images enables objective evaluation
of global as well as regional LV filling properties. As such, this
methodology promises to be valuable in the clinical assessment of
diastolic function in a large variety of heart diseases and
in the evaluation of therapeutic interventions on regional LV
diastolic properties.
Acknowledgments
Dr Philippe Vignon is supported by the French
Ministère des Affaires Etrangères (Fondation Lavoisier),
the Société Française de Cardiologie (Filiale
d'Echocardiographie), and Société de
Réanimation de Langue Française. We are indebted to our
sonographers, Beth Balasia, Jim Bednarz, Claudia Korcarz, and Joanne
Sandelski, for their contribution.
References
Top
Abstract
Introduction
Methods
Results
Discussion
References
This article has been cited by other articles:
 |
|
 |
|
  Y. Kato, K. Kotoh, A. Yamashita, H. Furuta, C. Shimazu, and T. Misaki Evaluation of Regional Aortic Distensibility Using Color Kinesis Angiology, May 1, 2003; 54(3): 345 - 351. [Abstract] [PDF]
|
|
|
|
 |
|
 V. Mor-Avi, K. A. Collins, C. E. Korcarz, M. Shah, K. T. Spencer, and R. M. Lang Detection of regional temporal abnormalities in left ventricular function during acute myocardial ischemia Am J Physiol Heart Circ Physiol, April 1, 2001; 280(4): H1770 - H1781. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Mor-Avi, K. Spencer, J. Gorcsan, A. Demaria, T. Kimball, M. Monaghan, J. Perez, J. P. Sun, L. Weinert, J. Bednarz, et al. Normal values of regional left ventricular endocardial motion: multicenter color kinesis study Am J Physiol Heart Circ Physiol, November 1, 2000; 279(5): H2464 - H2476. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Koch, R. M. Lang, M.-J. Garcia, L. Weinert, J. Bednarz, C. Korcarz, B. Coughlan, A. Spiegel, E. Kaji, K. T. Spencer, et al. Objective evaluation of regional left ventricular wall motion during dobutamine stress echocardiographic studies using segmental analysis of color kinesis images J. Am. Coll. Cardiol., August 1, 1999; 34(2): 409 - 419. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. VIGNON, L. WEINERT, V. MOR-AVI, K. T. SPENCER, J. BEDNARZ, and R. M. LANG Quantitative Assessment of Regional Right Ventricular Function with Color Kinesis Am. J. Respir. Crit. Care Med., June 1, 1999; 159(6): 1949 - 1959. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||