From the Division of Pediatric Cardiology, Department of Pediatrics
(M.A.F., P.M.W., J.R.), and the Department of Radiology (K.B.G., A.H., J.H.),
The University of Pennsylvania School of Medicine and The Children's
Hospital of Philadelphia, and the Department of Radiology (E.A.H.), University
of Iowa School of Medicine, Iowa City.
Correspondence to Mark A. Fogel, MD, Wyeth-Ayerst Research, Cardiovascular Division, C-2, 145 King of Prussia Road, Radnor, PA 19087. E-mail fogelm{at}war.wyeth.com
Methods and Results—MR myocardial tagging was used to examine 18
subjects with systemic LVs: 10 with functional single LVs (SLV) and 8
normal subjects (NL). Tracking the systolic motion of the
intersecting stripes were used to determine regional twist and radial
motion. Finite strain analysis was applied to derive principal
strains at the atrioventricular valve (AVV) and apical
short-axis levels and in 4 anatomic wall regions. Similar
E1 (circumferential shortening) strain and
heterogeneity of strain were noted between SLV and NL
except in the septal wall. At the septal wall, NL displayed greater
absolute strain (AVV=-0.16±0.02, apex=-0.17±0.02) and less
heterogeneity of strain than SLV (AVV=-0.12±0.02,
apex=-0.13±0.02). Similar E2 (wall
thickening) strain and heterogeneity of
strain were also noted between SLV and NL except again at the septal
wall. At the septal wall, SLV displayed greater absolute
E2 strain (AVV=0.17±0.08, apex=0.19±0.09)
and less heterogeneity of strain than NL
(AVV=0.07±0.07, apex=0.05±0.05). SLV twisted significantly less
counterclockwise than NL in 6 of 8 wall regions and actually twisted
clockwise at the AVV lateral wall. Although there was no significant
difference between groups in radial wall motion, the septal and
inferior walls of SLV demonstrated significantly less
radial motion compared with other SLV walls.
Conclusions—A major influence of the RV on systemic LV strain and
radial motion occurs in the septal wall, whereas absence of the RV
causes marked differences in LV twist. These findings may yield clues
to the long-term functioning of the SLV and be useful in determining
strategies for RV augmentation of LV function.
Understanding that
ventricular-ventricular interaction is an
integral part of cardiac mechanics, it follows that if a ventricle is
absent and therefore no ventricular-ventricular
interaction is occurring, cardiac mechanics may be affected. Patients
with functionally single LVs (such as tricuspid atresia) who
presently undergo staged surgical reconstruction culminating in the
Fontan procedure7 8 9 10 afford a unique opportunity
to study how the RV changes LV mechanics. Comparing the mechanics of
the systemic LV coupled and uncoupled to an RV tests the hypothesis
that the absence of ventricular-ventricular
interaction changes LV strain and wall motion. This comparison also may
give some insight into the causes of heart failure (eg, poor
systolic shortening, clinical congestive heart failure,
significant atrioventricular valve
regurgitation, high end-diastolic
pressures, and so forth), which occurs in some single LVs. We chose
patients with tricuspid atresia without transposition of the great
arteries {S, D, S} to compare with normal subjects because their
LVs have similar anatomic and electrical
morphology.11 12 13
We have previously used a magnetic resonance tagging technique called
SPAtial Modulation of Magnetization (SPAMM)14 15 16
to evaluate regional strain and wall motion in congenital heart
disease6 17 18 and, specifically,
ventricular-ventricular
interaction.6 Strain and its distribution across
the myocardial wall were demonstrated to be altered in the systemic RV
when a pulmonary LV was not coupled to
it.6 In addition, systemic RV twist (in an LV,
twist is counterclockwise in the short
axis18–20) and radial wall motion (in an LV,
radial wall motion is generally uniform across the short
axis18) were also shown to be altered when the LV
was not present.6 The present study was
undertaken to determine if this is the case for the systemic,
morphologic LV as well and to determine what changes in particular
occur. Normal human LVs studied in our laboratory were used as
reference.18
A retrospective review of the cardiac catheterization
data on single LV subjects was undertaken, of which 7 were available
for review. Hemodynamic data were acquired within
2.3±1.1 years from the date of MRI. Cardiac index,
ventricular end-diastolic pressure, superior
vena cava oxygen saturation, pulmonary and systemic pressures,
and vascular resistance were all acquired.
Control Population
Informed consent was obtained from all subjects. The human
investigations committee approved the study protocol on February 4,
1992. No patient had any arrhythmias that precluded study in
the scanner. All patients fasted 4 hours before MRI.
Magnetic Resonance Imaging
Myocardial Tagging
Image and Data Analysis
Wall Motion
where (Xn, Yn), and
(Xn+1, Yn+1) were the
coordinates of the centroid of the triangle at phase n and n+1,
respectively (Xc, Yc), were
coordinates of the ventricular cavity centroid, and
Twist was computed by finding the angle
Wall motion data was displayed graphically as in Figure 1
Homogeneous Finite Strain Analysis
Statistics
To obtain the homogeneity of strain within the region (ie, the
dispersion of strain within an entire wall region), the standard
deviation of all the strains within a given region was used, indexed to
the average, meaning the coefficient of variation:
Cardiac Catheterization
E1 Strain
Distribution of E1 Strain
Atrioventricular Valve to Apical Plane Strain Ratio
(Distribution of Strain Along the Long Axis of the Ventricle)
E2 Strain
Figure 4D
An alternate strain notation for E1 is
subtracting the absolute value of the strains presented in this
study from 1. Similarly, the alternate strain notation for
E2 is adding the strains presented
in this study to 1. If conservation of mass is preserved, the
product of E1 and
E2 should equal 1. The
Table
Wall Motion
Notice the uniformity of counterclockwise motion of the normal LV
across both short-axis planes. Also note the extent of the twist, and
contrast this with the motion of the single LV (Figure 5A
Radial Motion
In normal subjects, both LV effects on RV
function1 2 3 5 28 and RV effects on LV
function2 4 28 have been observed. These effects
are important in functional single ventricles because there is no
second ventricle to augment the function of the systemic ventricle. We
have noted that some patients with single RVs after the Fontan
procedure have presented with failing
ventricles,6 17 and the absence of
ventricular-ventricular interaction may play a
role. Indeed, our recent work that demonstrated an in vivo difference
in systolic mechanics between a systemic RV with and without a
pulmonary pumping LV6 lends credence to
this idea.
Patients with single LVs may also present in heart failure, even though
they may have a more suitable systemic pumping chamber than the single
RV (because of their LV morphology). It is unclear whether LV
morphology really plays a role. A further understanding of single LV
systolic mechanics is needed not only to complete our
investigation into ventricular-ventricular
interaction in single ventricles,6 but to design
better surgical and medical treatment strategies.
Strain
Because the single LV septum does not have an attached RV, it is a
"free wall" as well. It could be argued that its mechanics should
mimic free wall mechanics. Our data demonstrates that this does not
occur. The LV septal wall is morphologically different from the free
walls11 12 and this fact probably plays a role in
explaining the behavior of the septum.
It is interesting that E1 and
E2 strains are inversely related to each
other in the septal wall in the single LV when compared with normal
control LVs (E1 less in control LVs and
E2 greater in control LVs). We refer to the
way E1 and E2
change relative to each other between types of hearts (single LV versus
normal LV). It might be expected, however, that if the single LV
E1 is less in control LVs, then
E2 might also be less in control LVs. One
explanation may be that our measurements are in 2 dimensions, and the
difference is in the third dimension (base to apex dimension).
Alternatively, this may represent a difference in the material
properties of the myocardium (eg, the level of anisotropy
or the elastic modulus-the amount of deformation relative to the
stress) between single LVs and normal LVs. Relative amounts of septal
wall coronary blood flow between single LVs and normal
LVs30 may also play a role.
Additionally, this inverse relation between
E1 and E2
strains between single LVs and normal LVs may represent a
fundamental change in septal systolic mechanics from mostly
circumferential shortening to predominantly wall thickening. The septum
does not have RV pressure to buttress itself against during contraction
and therefore may bulge radially and do less circumferential
shortening. A difference in septal fiber orientation among the 2 groups
may also explain this. Finally, septal remodeling may have occurred,
changing the E1 and
E2 relation, and may represent the
most energy-efficient way for a single LV to pump blood.
E1 and E2
strain dispersion within a wall region was also significantly different
only at the septum. The strain dispersion of
E1 at both short-axis levels were
significantly higher in the single LV than in normal LVs, whereas
E2 dispersion was significantly less at
both levels. This implies a more disorderly pattern of circumferential
shortening and a more orderly pattern of myocardial wall thickening
than in normal LVs. This is consistent with the altered fiber
orientation or remodeling hypothesis mentioned
above31 32 (eg, fiber orientation altered to
optimize wall thickening).
As mentioned above, if conservation of mass is preserved, the
product of E1 and
E2 should equal 1 in the alternate notation
mentioned above. Our analysis demonstrates that this strain
product for septal walls of the single LV was significantly greater
than the corresponding walls of the normal LV. Changes in myocardial
blood volume to the septal wall of the single LV may play a role in
this phenomenon. In addition, single LV septal wall strain products
were not significantly different from 1. The inherent error in these
types of measurements and considering that these strain products
were only 3% >unity is further evidence that these values are no
different from 1.
Wall Motion
Single LVs twisted less counterclockwise at both short-axis levels than
did normal LVs. This decrease in twist is consistent with the
decreased cardiac index and wall hypokinesia we35
and others36 37 38 39 have observed after the Fontan
procedure in both single RVs and LVs. The second ventricle may
therefore augment the twist of the systemic ventricle as well.
Interestingly, single LV lateral wall twisted clockwise,
consistent with the abnormal wall motion that Akagi et
al39 found in their single LVs. This can be
explained by anatomic myocardial fiber orientation, which is a
left-handed helix in the epicardium and a right-handed helix in the
endocardium.31
No significant differences were noted in radial wall motion between
subjects with single LV and control subjects. However, comparisons
within the single LV group, at both short-axis levels, demonstrated
that septal and inferior walls had significantly less
radial motion when compared with other walls. Again, this would be
consistent with altered fiber orientation as noted above and
may be the most energy-efficient way for a single ventricle to
pump.
We have previously discussed our twist findings in the normal LV when
compared with other studies in the
literature.6 17 18 Briefly, there are conflicting
reports on the direction of twist at the LV short axis.
We18 20 and others19 40
believe that a counterclockwise twist is present when viewed apex
to base and this twist is uniform throughout the short axis.
Limitations
There are other limitations to this imaging approach. First, as we have
previously discussed,18 2-dimensional strain
analysis excludes the remaining components of the 3-dimensional
strain tensor.42 Solutions to 3-dimensional
strain analysis, however, require a priori knowledge of
the deformation to interpolate the data that, by its very nature, is
dependent on assumptions. This may introduce error in the
calculations.
Second, we note that homogeneous strain analysis
assumes that deformation is constant within a given triangular element.
Strains, however, have been known to have some transmural variation.
This therefore represents the limit of resolution on the strain
data.
Use of the single LV as a model of
ventricular-ventricular interaction has its
limitations in that these hearts have undergone cardiopulmonary
bypass. We would anticipate, however, that the altered strain and wall
motion abnormalities noted would be more diffuse if
cardiopulmonary bypass played a role. Ideally, the control
group would have been normal hearts in patients who have undergone
thoracotomy, pericardial stripping, and a cardiopulmonary
bypass run.
Ages for subjects with single LV and control subjects are significantly
different. Although quantitative changes may occur in subjects with
single LV as they grow, it is unlikely that the fundamental differences
in regional strain, twist, and radial motion we have observed will
change with time. Further, if age-dependent differences do exist, it is
likely to affect all walls, not just the septal wall as in our
study.
Finally, many of the steps in the data collection and analysis
were performed by the same person. This adds a bias to the data;
however, we do not believe this would have significantly changed our
results.
Conclusions
Received November 5, 1997;
revision received March 10, 1998;
accepted March 17, 1998.
© 1998 American Heart Association, Inc.
Clinical Investigation and Reports
Mechanics of the Single Left Ventricle
A Study in Ventricular-Ventricular Interaction II
Abstract
Top
Abstract
Introduction
Methods
Results
Discussion
References
Background—Left
ventricular (LV) effects on right ventricular
(RV) function are well known. Less is understood about the effect of
the RV on systemic LV mechanics. To determine this interaction, we
compared systemic LVs with and without an RV mechanically coupled
to them.
Key Words: contraction • ventricles • Fontan procedure • magnetic resonance imaging
Introduction
Top
Abstract
Introduction
Methods
Results
Discussion
References
Normally, left
ventricles (LVs) and right ventricles (RVs) affect each other's
mechanics1 2 3 4 5 by augmenting the function of the
other ventricle. LV effects on RV pressure generation have been
demonstrated in the human1 3 and
animal2 5 heart as well as RV effects on LV
function.2 4 We have also recently described the
effects of a pulmonary LV on systemic RV
mechanics.6 The mechanism of this
ventricular-ventricular interaction is thought
to be the mechanical coupling of the ventricles through the
interventricular septum.3 Because it
is known that the myocardium constituting the ventricles
forms an anatomic continuum around these chambers, others have
suggested that the free walls might also affect the contralateral
ventricle independent of septal effects.2
Methods
Top
Abstract
Introduction
Methods
Results
Discussion
References
Patients
Eighteen subjects with systemic LVs were prospectively studied.
Ten subjects had a functional single LV (all had tricuspid atresia
{S, D, S}, Edward's classification IA or IB). All had completed
the Fontan procedure7 8 9 10 with an atrial lateral
wall tunnel type total cavopulmonary connection and were
followed at The Children's Hospital of Philadelphia between December
30, 1993, and September 30, 1995. All patients were clinically well
from a cardiovascular standpoint. All underwent
echocardiography within 8 months of MRI and had
qualitatively good ventricular shortening. All had either
trace or no semilunar valve regurgitation. All values
are mean±SD. Single LV patients' ages ranged from 3 to 26 years
(10.6±7.7 years, median=8 years), heart rate was 88±21 bpm, and time
from operation was 5.8±4.5 years (median=5 years). Patients had to be
stable enough to undergo a 1 hour of MRI scan under sedation. All
patients were in normal sinus rhythm and had no evidence on surface ECG
of altered electrical activation.
The LVs of 8 healthy adult volunteers, used in a previous
investigation,15 were studied with magnetic
resonance tagging as a control group. Ages ranged from 21 to 35 years
(26.2±4.3 years), and heart rate was 70±15 bpm.
Studies were performed on a Siemens 1.5-T Magnetom SP-63. All
patients were monitored with pulse oxymetry, ECG, and by direct
visualization by video monitor. Our scanning protocol has been
previously described in great detail.6 17 18
Briefly, after localizers were performed, T1-weighted transverse images
were acquired throughout the thorax to evaluate
cardiovascular anatomy. Standardization and
localization of the short axis of the LV was performed by using the
transverse images. This was performed by choosing the long axis of the
LV as passing between the center of the mitral valve in the
anteroposterior plane (at the left edge of the aortic root) and the LV
apex. The short axis is perpendicular to this. Finally, myocardial
tagging with image acquisition was done. A high temporal resolution (20
ms) cine sequence was performed through the LV outflow tract to
determine the timing of end-systole, which was defined as aortic valve
closure. Two short-axis levels (one was one third of the way from the
atrioventricular valve to the apex [designated as
"atrioventricular valve"] and one was two thirds
of the way from the atrioventricular valve to the apex
[designated as "apex"]) were chosen for SPAMM imaging.
The SPAMM sequence has been previously
described.14 15 16 Briefly, SPAMM imaging uses a
prepulse sequence, applied immediately after the R wave, that saturates
2 series of parallel stripes perpendicular to each other. Each set of
parallel stripes is generated by 5 radiofrequency pulses separated by
field gradients. The tissue is thus divided into "cubes of
magnetization," as we have chosen to call
them6 17 18 (because each image has thickness)
(Figure 1
). In-plane cardiac motion moves
and distorts these "cubes of magnetization." Tracking this movement
and distortion enables measurement of strain and wall motion. We
acquired 6 to 8 gradient-echo images throughout systole starting
immediately after the SPAMM prepulse with the following
parameters: the repetition time (TR) was the RR interval
(range 650 to 1100 ms), flip angle=30 degrees, thickness=6 to 10
mm, inversion time (TI)=16 ms, number of excitations=3 and matrix
size=128x256 interpolated to 256x256, and the field of view ranged
from 200 to 400 mm. End-diastole was determined by the
R wave on ECG. The separation of the grid lines was selected to allow 3
to 4 lines between endocardial and epicardial surfaces (ie, 3 to 4 rows
of cubes). Black band (tag) width ranged from 1.5 to 2 mm.
View larger version (72K):
[in a new window]
Figure 1. SPAtial Modulation of Magnetization (SPAMM)
imaging and image analysis. The right of the image is anterior
and rightward, the left is posterior and leftward, the top is superior,
and the bottom is inferior. SPAMM intersection points were
tracked from end diastole (upper left) to end systole
(upper right). The centroid of each triangle was used to compute
regional wall motion (lower right). Dots represent
end-diastolic triangle centroid location and tails
represent subsequent motion. We computed 2-dimensional finite
strains by measuring triangle deformation. Data are coded in gray scale
(middle panel is the E2 strain map) and
superposed onto the anatomic image (lower left). White areas
represent regions of most wall thickening; darker gray and
black areas are regions of least wall thickening.
Images were transferred to a Sun SPARC 10 workstation (Sun
Microsystems). The VIDA (Volumetric Image Display and
Analysis)21 software package was used to
manipulate the images and perform all measurements as described
previously.6 17 18 Evaluation of wall motion and
strain has also been previously described.6 17 18
The appendix in our previous publication details the mathematics of
strain calculations.17 In brief, the first step
was to track the magnetically tagged grid intersections through systole
on a computer-based image by a video cursor and mouse with image
processing software.21 Delaunay
triangulation22 23 was used to create a
triangular grid automatically from the intersections that provided
uniform, nonoverlapping triangles. To compute regional wall motion
using wall motion software on the computer, the centroid of each
triangle was used. The regional deformations of the
myocardium were then characterized by using
homogeneous finite strain analysis on the deforming
triangles6 17 18 24 25 26 27 with the strain software
module of VIDA.21 This methodology has been
validated in a phantom14 and used in vivo by
Young et al.27 These studies demonstrated that
homogeneous strain analysis produced unbiased
estimates of the principal strains, principal angles, and orientations
of the principal axes.
As described previously,6 17 18 after
computing the centroid of all triangles during each systolic
phase, the (X, Y) coordinates of the ventricular
cavity centroid were obtained on the basis of the endocardial border
(Figure 1
). Twist and radial shortening were then measured by using the
relative motion of the centroid of the triangles relative to the
ventricular cavity centroid. How far the muscle marked by
the triangle was displaced away from the ventricular cavity
centroid (radial motion) from phase n to phase n+1 is described by
Equation 1:
n and
n+1 were lengths of the vectors from
the ventricular cavity centroid to the centroid of triangle
at phase n and n+1, respectively. Radial motion was taken as the net
inward motion of the centroid of each triangle toward the
ventricular cavity centroid relative to the
end-diastolic distance (measured in pixels of distance
moved divided by initial radial length to normalize for heart size). By
convention, radial motion inward was positive, and outward was
negative. 
made by 2 vectors drawn
from the ventricular cavity centroid to the centroid of the
triangle at phases n and n+1 as described in Equation 2:
where
n and
n+1 were vectors
(n and
n+1 were the length of those vectors)
from the ventricular cavity centroid to the centroid of
triangle at phase n and n+1, respectively, and
n · n+1 was
the vector dot product. Net twisting (twist measured from
end-diastole to end-systole, sampling at 7 time points) was
quantified for magnitude and direction. By convention, clockwise motion
was negative and counterclockwise was positive, viewing the heart from
apex to base. . "Dots"
are the location of the centroid of the triangle at
end-diastole and "tails" represent the
subsequent motion. The myocardial wall was divided into standard
anatomic regions (septal, inferior, lateral, and anterior
walls) by using the papillary muscles and other anatomic landmarks from
the transverse images as reference to perform the analysis.
Details of the mathematics are outlined in the appendix in our
previous investigation.17 Briefly, Delaunay
triangulation of the tagged intersections was used to compute uniform
triangles across the wall.22 23 Continuum
mechanics mathematics, which characterized the complex deformation
patterns of the myocardium,25 has
been described and used by us6 17 18 and
others,27 by using homogeneous finite
strains to characterize 2-dimensional shape changes of the magnetically
tagged grids. As we have noted, this approach assumed that deformations
within each triangle relative to end-diastole were locally
homogeneous (Figure 1). After a Lagrangian (Green's)
strain tensor, E was computed for each triangle; the strain tensors
were diagonalized so that they were independent of any coordinate
system. The local deformations were described by 2 principal strains,
Ei, and the orientation of the principal
axes relative to the original coordinate system. The first principal
strain, E1, was defined as the most
negative strain (also called compressive, circumferential shortening,
or minimum principal strain). The second principal strain,
E2, which was orthogonal to
E1, was defined as the most positive strain
(also called thickening, stretch, or maximal principal strain). Strains
reported in this study were obtained by averaging the strain of all the
triangles within the region. Strains are reported as mean±SD. Our
research focused on the maximum average systolic strain in each
region, which occurred at end-systole. The data were quantified and
displayed for qualitative analysis in color-coded or gray scale
form superimposed onto the anatomic images (Figure 1).
Comparisons between 2 means or a mean with a hypothesized value
were made by use of the unpaired, 2-way or 1-way (where appropriate)
Student's t test and the Wilcoxon ranked sum test.
Differences between various groups of subjects and location (regional
wall location, short-axis slice of either apex or base) were
analyzed by 2-factor ANOVA with repeated measures used when
appropriate. Comparison between multiple means within groups was done
with a 1-way ANOVA with pairwise comparisons made by use of Dunnett's
test or the Tukey-Kramer honestly significant difference test. All
measurements are mean±SD. Intraobserver variability was determined by
replicate measures and used the Student's t test. A single,
trained observer performed all image analysis steps.
Significance was defined as P<0.05. Statistical
analysis was performed on a personal computer with JMP version
3.1.4 (SAS Institute).
To compare strains between atrioventricular
valve and apical short-axis planes, the natural logarithm (ln) of the
ratio of the atrioventricular valve to apical strain
(eg, ln [atrioventricular valve strain/apical
strain]) was used. Besides stabilizing the variances in the data, this
was advantageous in that significant differences implied significant
differences in the geometric means of the two groups being
compared.
(3)
Results
Top
Abstract
Introduction
Methods
Results
Discussion
References
In each category (strain and wall motion), data are grouped
according to short-axis level (atrioventricular valve
and apical) and of course, subject group.
Cardiac index was 2.7±0.6 L ·
min-1 · m-2, LV
end-diastolic pressure was 9.4±3.3 mm Hg, aortic
pressures were 115±14.1/70±10.6 mm Hg, systemic vascular
resistance was 30.5±11.4 Wood units/m2, and
superior vena cava oxygen saturation was 62.2±6.1%. Mean
pulmonary artery pressure was 12.2±2.6 mm Hg, and
pulmonary vascular resistance was 1.8±0.8 Wood
units/m2.
Figure 2A displays strain data
in gray scale form superimposed on the anatomic image. Figure 2B and 2C
display principal compressive strain (E1
multiplied by -1) for both single LV and normal subjects in the 4
anatomic quadrants (anterior, inferior, posterior, and
superior walls) evaluated in 2 short-axis planes (1 near the
atrioventricular valve and 1 near the apex).
Coefficient of variation for strain measurements was 5.3±2.1%. At
both the atrioventricular valve (Figure 2B) and apical
levels (Figure 2C) single LVs demonstrated significantly less strain
than the control group only at the septal wall (P<0.05).
Furthermore, in comparing strains within groups at both levels, single
LVs again demonstrated significantly less strain in the septal wall
when compared with other wall regions (P<0.05), whereas
there was no significant difference among wall regions in the control
group.
View larger version (46K):
[in a new window]
Figure 2. E1 strain (A)
mapped onto the anatomic image of normal and single left ventricles
(LV) and displayed in B and C in bar graph format. A, The
atrioventricular valve (AVV) short-axis plane is on top
and the apical short-axis plane (APEX) is on the bottom of normal (left
side) and single (middle) LVs. Orientation is as in Figure 1 . The
E1 strain map is shown on the extreme right.
Note the strain difference in gray scale at the septal wall of single
LV with the other wall regions and normal LVs. B and C,
E1 strain (multiplied by -1) divided by
wall region and subject type at the AVV (B) and apical (C) short-axis
levels for normal LVs (NL) and single LVs (SLV). Error bars
represent SD. *Significantly less septal wall strain than
control group and within group at both levels.
Heterogeneity of Strain Within a Given
Anatomic Region
Similar to the absolute measures of strain above, at both the
atrioventricular valve (Figure 3A) and apical (Figure 3B)
levels, it was only the septal wall of single LVs that
differed significantly from controls, displaying greater
heterogeneity of strain (P<0.05).
Similarly, when comparing wall regions within each group, it is solely
the septal wall of single LV subjects that demonstrated a greater
heterogeneity of strain than other wall regions
(P<0.05), whereas no significant difference was noted in
normal subjects.
View larger version (33K):
[in a new window]
Figure 3. Distribution of E1
strain. Coefficient of variability (multiplied by -1) for
E1 strains (heterogeneity of
strain) by wall regions and group at the
atrioventricular valve (A) and apical (B) short-axis
levels. Error bars represent SD. SLV indicates single left
ventricle; NL, normal. *Significantly greater coefficient of
variability (heterogeneity of strain) than the control
group and within group only at the septal wall.
No significant differences were noted between single LV and normal
groups or when comparing wall regions within each group.
Figure 4A displays
E2 strain data in gray scale form
superimposed on the anatomic image and Figure 4B and 4C display the
principal E2 strain in graphical form. As
with E1 strain, at both the
atrioventricular valve (Figure 4B) and apical levels
(Figure 4C), single LVs demonstrated altered strain at only the septal
wall when compared with control. Single LVs displayed significantly
more E2 strain (myocardial wall thickening)
than the control (P<0.05). However, when comparing wall
regions within groups, no significant differences were noted.
View larger version (61K):
[in a new window]
Figure 4. E2strain mapped onto the anatomic image of normal and single left
ventricles (LV) and displayed in bar graph format. A,
Atrioventricular valve (AVV) short-axis plane is on top
and apical short-axis plane (APEX) is on bottom of normal (left side)
and single (middle) LVs. Image orientation is as in Figure 1 . The
E2 strain map is shown on the extreme right.
Note the strain difference in gray scale at the septal wall of single
LV with the other wall regions and normal LVs. B and C,
E2 strain divided by wall region and subject
type at the AVV (B) and apical (C) short-axis levels for normal LVs
(NL) and single LVs (SLV). Error bars represent SD.
*Significantly greater strain than control group only at the septal
wall. D and E, Coefficient of variability for
E2 strains (heterogeneity of
strain) by wall region and group at the
atrioventricular valve (D) and apical (E) short-axis
level. Error bars represent SD. *Significantly less coefficient
of variability (heterogeneity of strain) than control
group only at the septal wall. and 4E display the heterogeneity of
E2 strain within a given anatomic region.
As with the absolute measures of E2 strain,
at both the atrioventricular valve (Figure 4E) and
apical levels (Figure 4F), single LV subjects demonstrated altered
strain at only the septal wall when compared with control. Single LV
subjects displayed significantly less heterogeneity of
E2 strain (myocardial wall thickening) than
the control group (P<0.05). However, when comparing wall
regions within groups, no significant differences were noted. displays the product of
E1 and E2 for
all anatomic wall regions in each patient group. Strain products of
all wall regions except the septal walls of the single LV are <1. The
strain products of the septal walls of the single LV were only 3%
greater than unity, and considering the error in the measurement could
well be below 1. A 1-way t test demonstrated that the strain
products of the septal walls of the single LV were not
significantly different from 1, although they were significantly
different from the septal walls of the normal LV (1.03 versus 0.91,
respectively, for the AVV, P=0.016 and 1.03 versus 0.88,
respectively, for the apex, P=0.001).
View this table:
[in a new window]
Table 1. Product of E1 and
E2 for All Anatomic Wall Regions in Single and
Normal LVs
Twist
Figure 5A depicts the twisting
motion in the short axis of single LVs and controls in graphical
format; Figure 5B and 5C display the data quantitatively. Intraobserver
variability=6.7±2.1%. At the atrioventricular valve
level (Figure 5B), single LV subjects twisted significantly less
counterclockwise in 3 of 4 wall regions (except for the
inferior wall) when compared with controls and at the
lateral wall, twisted clockwise. When comparing wall regions within
groups, the lateral wall of single LVs (twisted clockwise) differed
significantly from the other wall regions that twisted counterclockwise
(P<0.05). No significant differences between wall regions
were noted in the control group. At the apical level (Figure 5C),
similar to the atrioventricular valve level, single LVs
twisted significantly less counterclockwise in 3 of 4 wall regions
(with the exception at this level, of the anterior wall) when compared
with controls. When comparing wall regions within groups, no
significant differences were noted between wall regions in both subject
groups.
View larger version (47K):
[in a new window]
Figure 5. Twist and radial wall motion of normal and
single left ventricles (LV) displayed graphically (A) and in
quantitative (B through E) format. A, Wall motion plots at the
atrioventricular valve (AVV) short-axis plane (top) and
apical (APEX) short-axis plane (bottom) of normal (left side) and
single LVs (middle). Orientation is as image orientation in Figure 1 .
Dots represent end-diastolic triangle
centroid location; tails represent subsequent motion. Note the
uniform counterclockwise twist and radial motion of normal LVs compared
with single LVs. Observe the clockwise twist of the single LV at the
lateral wall at the AVV plane. ED indicates end diastole.
B and C, Net twist by wall region and subject type at
the AVV (B) and apical (C) short-axis levels. Values are in degrees.
Positive values=counterclockwise motion, negative values=clockwise
motion. Error bars represent SD. SLV indicates single left
ventricle; NL, normal LV. *Significantly less counterclockwise twist
compared with control LVs. Note clockwise twist at the lateral wall at
the AVV level. D and E, Net radial motion by wall region and
group at the AVV (D) and apical (E) short-axis levels. Values are in
millimeters of movement indexed to end-diastolic radius
in mm (mm ED rad). Positive values=motion inward, negative
values=motion away (paradoxical motion). Error bars represent
SD. *Within-group comparison reveals significantly less radial motion
compared with lateral and anterior walls.). Observe the
clockwise motion of the lateral wall of the single LV at the
atrioventricular valve plane.
Figure 5A also depicts the radial motion in the short axis of
single LVs and control LVs in graphical format; Figure 5D and 5E
display the data quantitatively. Intraobserver variability=6.4±2.2%.
At both the atrioventricular valve (Figure 5D) and
apical levels (Figure 5E), no significant differences were noted
between single LVs and normal control LVs. When comparing wall regions
within groups, at both short-axis levels, the septal and
inferior walls of single LVs demonstrated significantly
less radial motion when compared with the lateral and anterior walls
(P<0.05). In control LVs, no significant differences were
noted among the wall regions.
Discussion
Top
Abstract
Introduction
Methods
Results
Discussion
References
Some authors have suggested that
ventricular-ventricular interaction is caused
by the shared septal wall, while others have suggested that the free
walls affect the contralateral ventricle independently of the
septum.1 2 3 4 5 6 28 The mechanism of free wall
contribution is the fact that the myocardium is a
"syncytium of muscle fibers tethered within a collagen
network"28 with muscle fibers in continuity
between RV and LV.29 Santamore et
al2 showed in rabbits that the free wall of one
ventricle affects the contralateral ventricle.
This study demonstrated that the single LV septal wall had a
significantly lower "circumferential shortening" strain
(E1) and a significantly higher
"stretch" E2 strain than in normal
subjects. This was the only wall region that differed from the control
group. Therefore, at least from absolute strain measures, it is the
septal wall that is affected the most by the RV. This may either be a
direct effect of RV pressure on the septal wall, or it may simply be a
"passive" effect; the absence of the RV as an anchor for the LV to
pump blood.
As we and others have noted,6 17 18 33 LV
systolic twist plays a role by distributing stress, strain, and
energy requirements across the ventricle. It enhances
diastolic filling by storage of potential energy and
diastolic elastic
recoil.6 17 18 33 34 This wall motion is thought
to be a result of complex fiber architecture31 32
and electrical activation. Altered twist may effect all parts of the
cardiac cycle with detrimental effects on function.
We have previously discussed6 17 18 in great
detail the limitation of acquiring data in fixed planes as
2-dimensional MRI does, with possible artifactual deformations caused
by motion into and out of this plane.41 We have
discussed6 17 18 in detail the study of Moore et
al41 and noted that we studied both single LV and
control subjects by the same protocol. One must remember that because
our study deals with large thickness slices relative to the amount of
through plane motion, it is not thought that this would have an
appreciable effect on the findings.
It appears that the RV affects systemic LV mechanics at the septal
wall only by decreasing the absolute circumferential shortening strain,
increasing the wall thickening strain, and increasing general strain
dispersion within a region. However, a decrease in the relative
endocardial strain with respect to epicardial strain was noted in the
lateral wall in the absence of an RV. Furthermore, a significant
decrease in LV twist relative to normal subjects was noted in patients
with single LV, with reversal of the normal counterclockwise rotation
at the lateral wall. Ventricular-ventricular
interaction, which has been known to occur in the normal human heart
and the systemic RV, is evident in vivo by looking at the effects of
the absence of an RV on LV wall mechanics. This study adds to the body
of knowledge on ventricular-ventricular
interaction and single ventricle mechanics and may lead to a better
understanding of LV function. The knowledge may contribute to designing
better medical and surgical treatments for the patient with single
ventricle.
Acknowledgments
Dr Fogel was funded through a fellowship grant of the
Southeastern Pennsylvania affiliate of the American Heart Association.
Dr Hoffman is an established investigator of the American Heart
Association.
Footnotes
Presented in part at the 69th Scientific Sessions of the American Heart Association in New Orleans, La, November 10–13, 1996, and previously published in abstract form (Circulation. 1996;94[suppl I]:I-538).
References
Top
Abstract
Introduction
Methods
Results
Discussion
References
This article has been cited by other articles:
 |
|
 |
|
  R. G. Williams, G. D. Pearson, R. J. Barst, J. S. Child, P. del Nido, W. M. Gersony, K. S. Kuehl, M. J. Landzberg, M. Myerson, S. R. Neish, et al. Report of the National Heart, Lung, and Blood Institute Working Group on Research in Adult Congenital Heart Disease J. Am. Coll. Cardiol., February 21, 2006; 47(4): 701 - 707. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Tanoue, H. Kado, T. Maeda, Y. Shiokawa, N. Fusazaki, and S. Ishikawa Left ventricular performance of pulmonary atresia with intact ventricular septum after right heart bypass surgery J. Thorac. Cardiovasc. Surg., November 1, 2004; 128(5): 710 - 717. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D J Sahn and G W Vick III Review of new techniques in echocardiography and magnetic resonance imaging as applied to patients with congenital heart disease Heart, December 1, 2001; 86(90002): ii41 - 53. [Full Text] [PDF]
|
|
|
|
|
|
K. Niwa, J. K. Perloff, S. Kaplan, J. S. Child, and P. D. Miner Eisenmenger syndrome in adults: Ventricular septal defect, truncus arteriosus, univentricular heart J. Am. Coll. Cardiol., July 1, 1999; 34(1): 223 - 232. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||