| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1995;60:986-997
© 1995 The Society of Thoracic Surgeons
Division of Cardiothoracic Surgery, Department of Surgery, and Department of Mechanical Engineering, Washington University, St. Louis, Missouri
 |
Abstract |
|---|
 TopFootnotes Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Methods. Transmural infarcts were created in adult Dorsett sheep (n = 8) by ligation of the distal homonymous coronary artery and were allowed to mature into left ventricular aneurysms for 8 to 12 weeks. Animals were imaged subsequently using double oblique magnetic resonance imaging with radiofrequency tissue tagging. Short axis slices were selected for analysis that included predominantly the septal component of the aneurysm as well as adjacent BZ regions in the anterior and posterior ventricular walls. Dark grid patterns of magnetic presaturations were placed on the myocardium and tracked as they deformed during the diastolic, isovolumic systolic, and systolic ejection phases of the cardiac cycle. Regional ventricular wall strains were calculated in BZ regions and regions remote from the aneurysm and compared with strains measured in corresponding regions from normal control sheep (n = 6).
Results. Diastolic midwall circumferential strains (fiber extensions) were relatively preserved, but abnormal circumferential lengthening strains were observed in the BZ regions during isovolumic systole. Peak circumferential strains ranged from 0.04 to 0.07 in the BZ regions but averaged -0.05 in the normal hearts (p = 0.002 for the anterior BZ and p = 0.001 for the posterior BZ). Midwall end-systolic fiber strains were depressed in the anterior BZ (-0.03 to -0.09 for the BZ versus -0.11 for the normal heart, p< 0.0001) but not in the posterior BZ (p = 0.19).
Conclusions. Our data support the theory that the stretching of BZ fibers during isovolumic systole contributed to a reduction in fiber shortening during systolic ejection and thus reduced the overall contribution of these fibers to forward ventricular output.
|
Introduction |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Left ventricular aneurysm (LVA) is a significant complication of myocardial infarction that may lead to global left ventricular (LV) dysfunction, ventricular arrhythmias, or thromboembolic complications [1]. For acutely ischemic myocardium, the pathophysiology of the global LV dysfunction has been linked to regional dysfunction in the border zone (BZ) region of normally perfused but poorly functioning myocardium [2]; however, the exact mechanism underlying the abnormal function has not been elucidated. Nicolosi and Spotnitz [3], using two-dimensional echocardiography to examine regional function in the BZ of LVAs, demonstrated a decrease in wall thickening in the BZ region during isovolumic systole but preserved overall systolic thickening. Lessick and colleagues [4] used cine-computed tomography in patients with LVA and demonstrated elevated wall stress and reduced wall thickening in the BZ region. Dor and colleagues [5] demonstrated improvements in systolic wall motion in the BZ and remote regions of myocardium in angiographic studies after surgical repair of LVA with a patch plasty reconstruction technique. These studies were limited, however, because cine-computed tomography, echocardiography, or angiography allowed only one-dimensional measurements of deformation. That is, only the radial component of deformation (wall thickness or wall motion) was measured, ignoring the circumferential, longitudinal, and shearing deformations. Furthermore, only the endocardium and epicardium could be tracked with cine-computed tomography or echocardiography so that the measurement of deformation was constrained to be constant transmurally across the ventricular wall.
Magnetic resonance imaging (MRI) tissue tagging [6] provides a noninvasive method for measuring the motion of ``material'' points in the myocardium and thus of determining transmural and locally varying deformation throughout the myocardium. Magnetic resonance imaging tissue tagging has been used to quantify two- or three-dimensional deformation in the normal [7, 8] and hypertrophic [9, 10] myocardium. These studies found regional variations in deformation in the normal heart and variations in several components of deformation in the hypertrophic heart.
The present study was undertaken to quantify the regional and time-varying pattern of deformation in an ovine model of LVA [11] and to compare it with corresponding regional deformations in normal animals to elucidate the mechanisms responsible for global LV dysfunction. Specifically, we sought to address the following questions: (1) Is the abnormality in systolic function in the BZ limited to the isovolumic phase of early systole as suggested by Nicolosi and Spotnitz [3]? (2) Are the relationships between strain components different in LVA versus the normal heart? For example, are shearing strains increased in the LVA border zone indicating a reduction in its contribution to the global LV deformation? (3) How does deformation in LVA vary with distance from the aneurysm and is deformation normal in regions remote from the aneurysm? We hypothesized that ventricular dysfunction in LVA may be attributed to abnormal circumferential lengthening strains in the LVA BZ during isovolumic systole that may lead to a reduction in systolic fiber contraction in the BZ regions. To answer these questions and investigate our hypothesis, we used a mathematical analysis technique [12] designed to extract maximal deformation data from the displacements of MRI tag lines and to quantitate myocardial strain in hearts with LVA.
|
Material and Methods |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
ANIMAL PROTOCOL FOR MAGNETIC RESONANCE IMAGING.
At 8 to 12 weeks after infarct, each animal was anesthetized with pentobarbital (30 mg/kg loading dose) and maintained on an infusion (80 to 100 mL/h) of guaifenesin (50 g/L), xylazine (100 mg/L), and ketamine (1,000 mg/L) and were ventilated with 100% oxygen using a high frequency jet ventilator (Mark 2; Bird Space Technology, Palm Springs, CA) to minimize respiratory motion. The first two animals in the series underwent LV ventriculography to verify the presence of LVA before MRI (Fig 1
) and subsequent animals had LVAs verified by postmortem autopsy. Additional normal sheep (n = 6) were imaged as controls. Nonferromagnetic catheters were placed in the left and right ventricles (Mikro-Tip, 3F in the right ventricle and 5F in the left ventricle, model SPR-458; Millar Instruments Inc, Houston, TX) of each animal through the carotid artery and jugular vein, respectively. The animal was positioned supine in the magnet (Siemens Magnetom SP-4000; Siemens Medical Systems, Iselin, NJ) with its chest centered in a Helmholtz coil. After MRI, the animals were sacrificed using intravenous pentobarbital (120 mg/kg) and heparin (5,000 U) followed by a rapid infusion of KCl (80 mEq). The heart was excised and preserved in formalin for subsequent histologic analysis.
|
). In all animals, a short axis slice was selected in which the aneurysm was located predominately in the septum and the BZ regions extended into the anterior and posterior ventricular walls (Fig 2B,C). For the selected short axis plane, a single slice MRI sequence consisted of a DANTE (delays alternating with nutations for tailored excitation) radiofrequency tissue-tagging preparation [13] followed by a two-dimensional FLASH (fast low flip angle shot) cine component. Imaging parameters included a slice thickness of 7 to 8 mm, field of view of 260 to 300 mm, 256 x 256 pixel matrix size, two to three acquisitions, 40° flip angle, echo time of 25 to 30 ms; time to recovery = one cardiac cycle; and intertag spacing = 5 to 6 mm.
|
) and the deformation of these presaturations or tags were used to calculate two-dimensional strains by a method described below.
Hemodynamic Data Acquisition and Analysis
During the scanning interval, the LV pressure, the first derivative of LV pressure and the trigger signal were recorded continuously using customized data acquisition and manipulation software (LabView 2, National Instruments, Austin, TX). The images resulting from a single experiment represented an average over many heart beats. Therefore, a signal averaging algorithm was performed on the hemodynamic data set to yield two-beat average data for each hemodynamic parameter (Fig 3A). The algorithm identified the smallest trigger-to-trigger interval and computed the mean signal and standard deviation for each hemodynamic parameter during this interval.
|
). Specifically, strains were calculated for the ejection, isovolumic systolic, and diastolic phases. The isovolumic phase was defined from the LV pressure tracing: early isovolumic systole occurred when the LV pressure increased by 5% of its maximum value from its diastolic baseline and the end of isovolumic systole occurred when the LV pressure reached its plateau. The ejection phase started at the LV pressure plateau and ended at the rapid downslope of the LV pressure. The diastolic phase started at the lowest LV pressure and ended at the beginning of the isovolumic phase. These definitions facilitated comparison of the strains for any phase of the cardiac cycle among animals. Because each animal had a different heart rate and a different systolic or diastolic phase duration, it was necessary to subdivide the isovolumic and systolic phases into three equal intervals (Fig 3B) and the diastolic interval into two equal intervals.
Image Processing and Localization of Tag Lines
Raw 256 x 256 pixel images were transferred to a Silicon Graphics Iris Indigo workstation (Silicon Graphics, Mountain View, CA) and were converted from Siemens format to SGI format using custom software. Images were cropped and scaled by 300% and displayed with the tag lines at 45° to the x,y coordinate system established by the MRI scanner. Tag lines and intersection points of the tag lines were located using a semiautomated algorithm [12]. Briefly, each tag line was fit with a fifth order polynomial in cartesian coordinates and the fit was refined iteratively by sampling the discrete pixel density function perpendicular to the polynomial, fitting a gaussian curve to the pixel density function and adjusting the polynomial based on the mean of the gaussian.
Strain Calculations
The method for calculating strains has been described elsewhere [12]. A particular BZ or remote region of interest in the LV was identified (see Fig 2). Using the polynomial representation of the tag lines and the tag intersection points in that region, strains were calculated on a regional basis by defining a mapping from the region on the undeformed image to the region on the deformed image. The mapping was defined by determining a set of fitting coefficients for a ``surface polynomial'' that was used to represent the x or y component of displacement.
Before defining the mapping, the x and y components of displacement were determined separately for each tag line. For example, in representing the x component of displacement, (Ûx)k, along a tag line, the y component of displacement along that tag line was interpolated using a one-dimensional isoparametric interpolation of the measured y displacements of the intersection points lying on that tag line. The measured x displacements along the kth tag line were then determined by the difference in x coordinates between the fifth order polynomial representing the deformed tag line and that representing the undeformed tag line:
 | (1) |
 | (2) |
where 
i(t) were one-dimensional isoparametric basis functions whose order was determined by the number of intersection points on the tag line [12], t was the standard parametric coordinate after the one-dimensional mapping, 
yk was the interpolated y displacement of a point on the kth tag line, Ûyi were the measured y displacements of the intersection points on that tag line, NPTS was the number of tag intersection points that were on the tag line and fdef and fundef were the fifth order polynomial representations of the deformed and undeformed tag lines, respectively, where the x coordinate was expressed as a function of y. The procedure was reversed in estimating the y displacements of the tag lines.
For each element, a mapping was defined from the undeformed to the deformed state as:
 | (3) |
where X = (x,y) was a point in the undeformed region and x = (x`,y`) was the mapped point on the deformed region. The displacement components of the mapping were Ux = x` - x and Uy = y` - y.
Each component of displacement, Ux and Uy, was expressed as a linear combination of the product of two-dimensional polynomial basis functions and unknown coefficients:
 | (4) |
 | (5) |
where 
i were the polynomial basis functions (see [12]), the ai, bi were unknown coefficients, and m was the number of basis functions. To define the mapping, for example for the x component of displacement, we sought coefficients ai, bi that minimized an objective function comparing model and measured displacements:
 | (6) |
where the first term expressed the fidelity of the surface polynomial and the measured deformations of the tag lines and the second term was a smoothing term to control the trade-off between fidelity of model and measurements and the smoothness of the interpolating surface polynomial. In equation (6), Ux(x,y) could be thought of as a ``surface polynomial'' that interpolated the x component of displacement for any point x,y in the domain 
of the myocardial region. The Ck were the fifth order polynomial representations of the tag lines, p was a parameter used to control the contribution of the smoothing term to the objective function S, (Ûx)k was the parametric representation of the measured displacements on tag line Ck, and NL was the number of tag lines. Minimization of equation (6) was a linear fitting problem in the coefficients, ai, i = 1, 2, ..., m. Minimization of equation (6) was reduced to solving the set of linear equations:
 | (7) |
 | (8) |
 | (9) |
 | (10) |
as explained in [12] where ux = (a1, ..., am)T. An analogous fitting was performed to determine the coefficients bi that defined the mapping of the y displacements.
A continuous distribution of finite strains was obtained by calculating partial derivatives of the approximating surface polynomials with respect to the spatial coordinates and applying the standard (Lagrangian) definition of finite strains [15]. Circumferential and radial strains were calculated along the midwall curve (see Fig 2) by computing the tangent to the midwall curve. The strain tensor was rotated through the angle made by the tangent vector and the x-axis giving circumferential (ec) and radial (er) components of tensile strain and the radial-circumferential (erc) shear strain. Principal strain components (e1, e2) were calculated by solving an algebraic eigenvalue problem where, for a two-dimensional problem, ex and ey were the diagonal elements in the 2 x 2 symmetric matrix and exy was the off-diagonal element. Diagonalization of this 2 x 2 tensor representation of the strain components gave the principal strains. Physically, e1 represented the strain in the direction of greatest stretch (positive strain) and e2 represented the strain in the direction of greatest negative (or least positive) strain.
Definition of Border Zone and Remote Regions
The LV myocardium on the short axis images was subdivided into three regions-aneurysm, border zone, and remote regions. The anterior BZ (ABZ) was defined as the region of anterior wall myocardium adjacent to the aneurysm and the posterior BZ (PBZ) was defined similarly as the region adjacent to the posterior border of the aneurysm (see Fig 2). Corresponding regions were identified on the normal hearts. Using the data from Dor and colleagues [5], BZ regions were defined as beginning at the region of maximal curvature adjacent to the anterior and posterior extent of aneurysm and ending two tag boxes from this point (see Fig 2). Thus, the BZ regions extended two times the width of a tag box diagonal; that is, approximately 1.4 cm. The BZ regions were subdivided further into three regions of equal size (REG1, REG2, REG3). The remote region was defined as the region equidistant from the ABZ and PBZ. Strains were calculated as a continuous function using equations (7 to 11) along the midwall curve (see Fig 2). For each region (ABZ, PBZ, and remote), strains were computed from the continuous function at 21 equidistant points along the midwall curve spanning the entire distance of the region. These points were averaged to give a mean strain for that region. Thus, for each subregion of the ABZ or PBZ, seven points were averaged to compute a mean strain for that subregion. The mean strain ± standard deviation (first and second principal [e1, e2], radial [er], circumferential [ec], and radial-circumferential shear strain [erc]) was computed for each subregion by averaging the strains for that subregion from all the animals.
Histology
Quantitative histology was performed on four hearts with LVAs to assess collagen content and fiber architecture in the aneurysm, BZ, and remote myocardium. Quantitative estimates of collagen content were made with computer-assisted planimetry. Briefly, multiple sections of aneurysmal, BZ, and remote myocardium were cut from the fixed hearts, sectioned and stained with a Masson trichrome stain that stains collagen green and muscle red. Histologic slides were scanned into a 512 x 512 computer image. Images were analyzed using Adobe Photoshop 2.0.1 (Adobe Systems). Pixels with a green hue (collagen) were identified using Adobe and were converted to black; all other pixels were converted to white. Computer-assisted planimetry was performed using National Institutes of Health IMAGE (v. 1.54, Wayne Rasband, National Institutes of Health, Bethesda, MD). The number of black and white pixels were quantitated, and their relative area of contribution to the entire image was determined. Results were reported as the mean ± standard deviation of the percentage of the total pixels that were black (collagen).
Validation
The method of computing strains described in the Strain Calculations section was validated extensively using a finite element model in [12].
Statistics
Differences in strain components between remote regions of myocardium in the LVA and the normal heart were made with an independent t-test assuming unequal variances and comparisons among ABZ and PBZ regions were made with one-way analysis of variation and a Tukey test. Collagen content was evaluated with a t test.
|
Results |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Diastolic Strains
There were only small differences in end-diastolic strains between the ABZ or PBZ and the normal hearts. In particular, there was a trend toward a reduction in circumferential lengthening in the PBZ when compared with the corresponding region in the normal heart; however, this did not reach statistical significance (ec = 0.17 ± 0.12 in the normal heart and ec = 0.05 ± 0.04 in the PBZ; p = 0.19).
Isovolumic Strains
Significant differences existed between strain components in the ABZ and PBZ versus the normal heart during the isovolumic period. In particular, in both the ABZ and PBZ, there was abnormal isovolumic circumferential lengthening at all time intervals when compared with the normal heart (Fig 4A,B). In the PBZ, radial thinning was observed when compared with the normal heart (Fig 4B). Circumferential shortening was significantly greater in the normal hearts versus the remote region of the LVA hearts (Fig 4C).
|
Overall Function
Plots of cumulative strain over the period from early diastole (passive filling) to the end of systolic ejection are demonstrated for the BZ and remote regions in Figure 5.
|
|
Comment |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This study used the recently developed technique of MRI tissue tagging [6] to quantitate strains in LVA. In MRI tagging, dark grid patterns of magnetic presaturation are placed on the myocardium using radiofrequency pulses and changes in magnetization. Because these presaturations represent actual changes in the magnetization of protons in the myocardium, the tag lines deform as the heart deforms. Thus, displacements of tag lines on a two-dimensional image represent actual in-plane motion of ``material points'' in the heart. By using both tag lines and intersections of tag lines, we were able to interpolate the two-dimensional displacements of material points in the myocardium. The x and y components of displacement of these discrete points were then fit by a smooth function (the surface polynomial) giving a continuous functional representation of displacement in a region of myocardium. Using continuum mechanics, finite strains were computed from derivatives of the functional representation of the displacement field.
Myocardial wall strain represents a relative change in length between any two material points in the myocardium. That is, strain is defined as the difference between the length of a line segment connecting two points in the deformed configuration minus its length in the undeformed state divided by its undeformed length. Thus, wall thickening or fractional shortening are measures of strain. The strains computed in this study differ from conventional measures of strain in three significant ways: First, because the surface polynomial method computes a continuous functional description of displacement over a region of myocardium, the strains computed from the derivatives of this function represent strains between two material points separated by the infinitesimal distances dx and dy, whereas measurements of strain using wall thickening use a finite length, the wall thickness, between the material points. Thus, we were able to compute a continuously varying distribution of strain from one local area of myocardium to the next. Second, continuum mechanics allows one to calculate all components of strain at a particular point. In two dimensions, there are three components of deformation at each point: two tensile strains (for example, radial and circumferential) and one shear strain (for example, radial-circumferential shear strain). Wall thickening only measures one component of deformation (radial) and thus may miss significant alterations in function. For example, our data showed significant reduction in circumferential strain in the ABZ in systole but no change in the radial component of strain. This was the same conclusion reached by Nicolosi and Spotnitz [3]; however, based on their data they concluded that systolic contraction was preserved while we were able to examine deformation in a more detailed manner and to conclude that systolic contraction was abnormal in the ABZ.
Furthermore, because radial, circumferential, and shearing components of deformation during systole in the heart are generated by different mechanisms-radial thickening by rearrangement of fibers [16], circumferential strains by shortening of circumferentially oriented muscle fibers, and shearing deformations by the branching of muscle fibers [17]-the pathophysiology of conditions such as LVA can be elucidated only by examining all components of strain.
Finally, we calculated finite (nonlinear) strains that differed from the traditional infinitesimal (linear) definition of strain used in conventional descriptions of wall thickening by including second order terms in the definition of strain [15]. For soft tissues, such as myocardium, that undergo large strains (0.1 to 0.2), conventional linear definitions of strain may create errors when rigid body rotations are large.
Measurements of strain, wall thickening, or wall motion reflect the contraction (or stretching) of sarcomeres in that region. Thus, measurements of myocardial deformation provide information about the length component of the length-tension time model of cardiac contraction. Because muscle fibers are oriented predominantly in a circumferential and longitudinal direction with the longitudinal angle varying continuously through the ventricular wall [18], measurements of radial deformation reflect sarcomere shortening only indirectly. At the midwall of the LV, muscle fibers are oriented almost exclusively in a circumferential orientation with only a minimal longitudinal angle. Thus, midwall circumferential strains as measured in this study closely reflected the change in length of sarcomeres in the midwall of the LV. As alluded to previously, radial thickening of the myocardium in systole is postulated to occur by a complete three-dimensional rearrangement of muscle fiber bundles [16], and thus does not closely reflect the contraction of sarcomeres in these muscle bundles. Other measures of systolic deformation such as quantitative wall motion estimates or wall motion scores, while useful as qualitative measures of abnormal function, provide little insight into the functioning of underlying sarcomeres because the same systolic strain may occur with a myocardial displacement of, for example, 0.1 or 1 cm. The additional displacement in the latter case may be due to rigid body motion of the myocardium.
Our data demonstrate that diastolic strains are similar in the BZ regions and corresponding regions in the normal heart. In particular, while there was some heterogeneity of function, the midwall circumferential strains in the ABZ were not significantly different than the normal heart. In the PBZ, the circumferential strains did not differ significantly from the normal heart; however, there was a tendency toward a reduction in circumferential extensions. These results were not consistent with the model predictions of Bogen and colleagues [19] or Janz and Waldron [20], who predicted using simple mathematical models of the heart that the stiff aneurysm would impede the extension of muscle fibers in the BZ regions. Our results indicated that there may be a small decrement in diastolic fiber extensions in the BZ, but that these differences were not of the magnitude predicted by the models. Possible explanations for the discrepancy between model and measurements is the relative simplicity of the models-both models ignored ventricular anisotropy with respect to the fibers-or an overestimation of the stiffness of the aneurysm in the models.
The principal abnormality in function demonstrated in this study was the lengthening (positive) midwall circumferential strains in isovolumic systole in both the ABZ and the PBZ compared with both the normal heart and remote regions of the aneurysmal hearts (see Fig 4A,B). These data confirmed the data of Nicolosi and Spotnitz [3] who measured wall thickening using echocardiography in patients before and after undergoing repair of LVA. They demonstrated abnormal wall thinning (radial thinning) in the BZ regions but preserved overall systolic thickening. The isovolumic lengthening may have contributed to systolic dysfunction by reducing the contribution of the BZ regions to overall systolic contraction. Midwall isovolumic circumferential strains in the remote regions of the LVA were increased significantly compared with corresponding regions in the normal heart; however, the strains were very close to zero in the remote region and were slightly negative in the normal heart. Because very small circumferential strains in the remote regions were expected during isovolumic systole, the difference may have reflected only a slight difference in timing of the isovolumic interval between the normal and aneurysmal sheep.
The abnormal isovolumic strains contributed to a reduction in midwall circumferential strains during systolic ejection in the ABZ. Although there was a tendency toward a reduction in negative systolic ec in the PBZ, it did not reach statistical significance. The systolic dysfunction in the ABZ extended only into REG1 and REG2, with REG3 displaying normal systolic function. Thus, the systolic dysfunction extended approximately 1 cm from the aneurysm rim. Interestingly, midwall radial strains were similar between BZ regions and the normal heart as indicated by Nicolosi and Spotnitz [3]. Thus, the abnormal systolic function was evident only when measuring both circumferential and radial components of deformation. In addition to abnormalities in midwall ec, shear strains were negative in the normal heart, but positive in the ABZ. Shear strains in the normal heart are caused by the branching of muscle fibers [17] and the helical arrangement of muscle fibers [18], which are disrupted adjacent to the BZ. Consequently, the opposition of normally contracting myocardium on one side of BZ and outwardly displacing aneurysm on the other may lead to a change in sign of shear strains. The absolute magnitudes of the shear strains were not increased. Thus, excessive shear strains, which would not be expected to contribute to useful myocardial work, cannot be invoked to explain global ventricular dysfunction.
Lessick and colleagues [4], using one-dimensional measurements of wall thickness in patients with LVA, indicated that the mechanical dysfunction in the BZ, manifested by reduced systolic wall thickening, extended 3 to 4 cm from the edge of the aneurysm. Conversely, in acutely ischemic myocardium Gallagher and co-workers [21] found only a narrow 1-cm region of normally perfused but poorly functioning myocardium, and Nanto and colleagues [22] found, using contrast echocardiography and wall thickening measurements, that the poorly functioning but normally perfused zone extended 1.3 cm from the ischemic zone in patients with acute ischemia. Our own results indicated that the ABZ region within 1 cm of the aneurysm edge (REG1 and REG2) displayed abnormal systolic circumferential shortening, whereas REG3 (from 1 to 1.4 cm from the aneurysm edge) had preserved systolic function. However, REG3 still had abnormal lengthening strains during isovolumic systole indicating that the BZ region extended at least 1.4 cm from the aneurysm edge.
Multiple theories have been proposed to explain global LV dysfunction in terms of dysfunction in the border zone. Some investigators have found increased wall stress in the BZ [5, 23] and attributed global LV dysfunction to the increased wall stress. Hutchins and Brawley [24] found a reduced curvature-thickness index in BZ regions of aneurysms in patients dying of LV dysfunction after aneurysm resection versus those dying of other causes. They speculated that an increased curvature-thickness index, which could be created by inverting the edges of the aneurysm during repair, would result in improved function. Other studies have speculated that mechanical tethering of the BZ region to adjacent noncontracting myocardium could contribute to global dysfunction [25]. Other possible explanations include a reduction in systolic blood flow or delayed activation of the BZ region.
No direct measurements of regional blood flow in the BZ of LVA have been performed to evaluate the significance of ischemia on BZ function. However, Cox and colleagues [26] demonstrated swelling of mitochondria and accumulation of neutral fat droplets in the BZ region surrounding an infarct, but overall preserved fiber architecture. These changes persisted at least 1 week after coronary ligation suggesting a degree of persistent ischemic injury. The histologic analysis in this study demonstrated an intermediate amount of collagen infiltration in the BZ compared with remote or aneurysmal regions (18.5% in the BZ, 1% in remote regions, and 90.5% in the aneurysm).
We performed electrophysiologic mapping of ventricular activation sequences in seven sheep with LVA using silicone epicardial plaque electrodes and a 256-channel recording system [27]. Two normal animals were also mapped. Average activation times in the BZ region were 9 ms and were 13 ms in normal animals. Thus, activation in the sheep was very rapid, in general, and was not reduced in the BZ region. Therefore, delayed activation did not contribute to the abnormal isovolumic lengthening seen in the BZ region.
Thus, the results of this study combined with the above information on the histology and cellular biochemistry of BZ muscle fibers imply that the overstretching of BZ fibers during isovolumic systole (Fig 6) may be explained by (1) an intrinsic abnormality of the BZ region and (2) a possible increase in wall stress. Border zone muscle fibers have been replaced by collagen and may have a persistent ischemic injury or may be hibernating myocardium. An increase in wall stress would be expected in BZ regions because they have undergone some wall thinning and because they are adjacent to a stiff aneurysm. The increase in wall stress would be greatest during isovolumic systole and coupled with the reduction in number of BZ fibers and a possible ischemic injury, the BZ fibers would be unable to develop sufficient isometric tension to counteract the increase in stress, thus resulting in stretching during isovolumic systole. Noma and colleagues [28] demonstrated that paradoxic isovolumic bulging in the BZ of acutely ischemic myocardium was proportional to afterload. The abnormal isovolumic stretching in the BZ contributed to a reduction in overall systolic fiber contraction in the BZ regions, which contributed to global LV systolic dysfunction.
|
) may eliminate the increased stress on BZ regions allowing the weakened region to undergo normal synchronized contraction. This technique may eliminate ``residual aneurysm'' after repair, which probably represents weakened, dyskinetic BZ regions. An important extension of our study then would be to examine the time-varying pattern of strains in the heart after repair of a LVA with a linear or circular technique.
|
An additional limitation was the limited transmural resolution of the tag lines. We were able to place tag lines separated by 5 to 6 mm. Smaller distances between tag lines resulted in rapid fading of the tag lines, thus limiting the transmural resolution to one or two tag boxes across the heart wall. Consequently, despite the fact that the transmural distribution of strain is of considerable interest, we limited our analysis to midwall strains that closely reflected fiber contractions in that region and we did not measure the transmural distribution of strains. It should be noted that although transmural resolution was limited, in contrast, the circumferential distribution of the tag lines was quite dense, thus allowing us to measure more precisely the regionally varying nature of midwall strains in LVA. Further improvements in MRI technology including improved sequence design, faster imaging methods, and better cardiac coils will be necessary to improve the resolution of measurements of transmural deformation.
In summary, MRI tissue tagging was used to calculate two-dimensional strains in the BZ and remote regions of an ovine model of LVA and were compared with corresponding regions of the normal heart. The ability to track multiple ``material points'' in the myocardium and the application of a sophisticated mathematical analysis technique allow a detailed examination of the time-varying, regional nature of the pathophysiologic deformation in LVA. Significant abnormalities were seen in isovolumic systole where the overstretching of BZ fibers contributed to reduced systolic fiber contraction in the BZ and thus may have explained partially global LV dysfunction in LVA. Furthermore, our results implied that a method of repairing an LVA that could prevent the isovolumic overstretching might allow preservation of systolic contractile function in the BZ regions.
|
Acknowledgments |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This project was supported in part by the National Institutes of Health under grants HL50139 and HL08746.
|
Footnotes |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Address reprint requests to Dr Pasque, Division of Cardiothoracic Surgery, Washington University School of Medicine, 3108 Queeny Tower, One Barnes Hospital Plaza, St. Louis, MO 63110.
|
References |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  P. Zhang, J. M. Guccione, S. I. Nicholas, J. C. Walker, P. C. Crawford, A. Shamal, G. Acevedo-Bolton, M. A. Guttman, C. Ozturk, E. R. McVeigh, et al. Endoventricular patch plasty for dyskinetic anteroapical left ventricular aneurysm increases systolic circumferential shortening in sheep. J. Thorac. Cardiovasc. Surg., October 1, 2007; 134(4): 1017 - 1024.e1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Langer, F. Rodriguez, A. Cheng, S. Ortiz, K. B. Harrington, M. K. Zasio, G. T. Daughters, J. C. Criscione, N. B. Ingels, and D. C. Miller Alterations in Lateral Left Ventricular Wall Transmural Strains During Acute Circumflex and Anterior Descending Coronary Occlusion Ann. Thorac. Surg., July 1, 2007; 84(1): 51 - 60. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. T. Wall, J. C. Walker, K. E. Healy, M. B. Ratcliffe, and J. M. Guccione Theoretical Impact of the Injection of Material Into the Myocardium: A Finite Element Model Simulation Circulation, December 12, 2006; 114(24): 2627 - 2635. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Cheng, T. C. Nguyen, M. Malinowski, F. Langer, D. Liang, G. T. Daughters, N. B. Ingels Jr, and D. C. Miller Passive Ventricular Constraint Prevents Transmural Shear Strain Progression in Left Ventricle Remodeling Circulation, July 4, 2006; 114(1_suppl): I-79 - I-86. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Cheng, T. C. Nguyen, M. Malinowski, D. Liang, G. T. Daughters, N. B. Ingels Jr, and D. C. Miller Effects of Undersized Mitral Annuloplasty on Regional Transmural Left Ventricular Wall Strains and Wall Thickening Mechanisms Circulation, July 4, 2006; 114(1_suppl): I-600 - I-609. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
The effect of anteroapical aneurysm plication on end-systolic three-dimensional strain in the sheep: a magnetic resonance imaging tagging study. J. Thorac. Cardiovasc. Surg., March 1, 2006; 131(3): 579 - 586.e3.
|
|
|
|
|
|
|
p = 0.04 for REG1 and p = 0.003 for REG2 in interval 2; 
p = 0.04 for REG2 in interval 3). In the posterior border zone, there is also significant early radial thinning (*p = 0.02 for REG1, p = 0.04 for REG2 in interval 1; 
0.
