Myxomatous mitral valve disease (MMVD) is the most common cardiac disease of dogs, and some dogs with MMVD develop myocardial dysfunction because of enlargement and remodeling of the heart. In humans, systolic dysfunction assessed by ejection fraction is associated with an increased risk of adverse events in patients with mitral regurgitation (MR), and this is crucial in deciding the optimal timing of surgery in patients with chronic severe MR. Moreover, dogs with moderate heart failure caused by MMVD also have decreased systolic function, which is associated with a decreased survival time. However, assessment of systolic function by conventional echocardiographic methods is difficult in MR owing to altered hemodynamic loading conditions.
Because left ventricular (LV) torsion is directly related to helically oriented myocardial fibers, it might provide better assessment of myocardial function than conventional methods and identify subclinical abnormalities in cardiac function.[6-8] LV torsional motion has been assessed invasively or with anesthesia such as with implanted radiopaque markers, sonomicrometry, and magnetic resonance imaging. Therefore, it is difficult to evaluate in clinical settings. Recently, two-dimensional speckle-tracking echocardiography (2D-STE) has been developed in which systolic LV torsion is assessed noninvasively, rapidly, and with no anesthesia.[8, 12, 13] 2D-STE tracks grayscale B-mode images of unique speckle patterns in the myocardium and therefore is independent of cardiac translation, tethering, and angle. Moreover, assessment of LV torsion in awake normal dogs and dogs with hypokinesia has been validated as a repeatable method for evaluating myocardial motion. However, LV torsion in the clinical setting of dogs with cardiac disease has not been used to evaluate their myocardial function.
To our knowledge, clinical assessment of torsional deformations by 2D-STE in dogs with MMVD has not been reported previously. Moreover, alterations of torsional deformations based on progression of MMVD have not evaluated. Therefore, the purpose of our study was to quantitatively measure systolic torsional deformations in dogs with MMVD and in weight- and age-matched controls.
Materials and Methods
- Top of page
- Materials and Methods
Our study population consisted of 67 client-owned dogs with MMVD and 16 weight- and age-matched healthy dogs serving as controls. These dogs were presented for cardiac screening at the Veterinary Medical Hospital of Nippon Veterinary and Life Science University in Japan and were analyzed retrospectively. Dogs that had already been treated by a referral hospital, except for dogs treated with pimobendan, were allowed into the study. The dogs were clinically diagnosed with MMVD by a previously reported echocardiography method. We included dogs with mitral valve prolapse, mitral valve leaflet thickening (as determined from 2D echocardiography), and MR (as determined from color Doppler examination).
We excluded dogs that presented with congenital heart disease or other acquired cardiovascular disorders. All dogs underwent a complete physical examination, electrocardiography, thoracic radiography, and transthoracic echocardiography. We excluded dogs with MMVD that also had a velocity of tricuspid regurgitant flow of >3.45 m/s, as determined by Doppler echocardiography. This regurgitation rate is equivalent to a peak pressure gradient of >48 mmHg, as calculated by the modified Bernoulli equation. In this study, given the influence of increased right ventricular pressure on LV motion, we excluded dogs with a right ventricular-to-right atrial systolic pressure gradient ≥48 mmHg that could not be improved by treatment that lowered left atrial load.
All 67 dogs with MMVD were placed into class I, II, or III according to the International Small Animal Cardiac Health Council (ISACHC) classification. Control dogs were healthy, with normal findings on complete physical examination, electrocardiography, thoracic radiography, and transthoracic echocardiography. None of the control dogs had a history of clinical signs of cardiac disease.
Conventional 2D, M-mode, and Doppler examinations were performed by a single trained investigator (H.K.) using an echocardiographic system1 and transducer,2 and simultaneous ECG limb lead II was recorded and displayed on the images. All data were obtained for at least 5 consecutive cardiac cycles in sinus rhythm, from nonsedated dogs that were manually restrained in right and left lateral recumbent positions. Images were analyzed by 1 of 4 trained observers using an off-line workstation.3 The presence of mitral valve prolapse, mitral valve thickening, and MR was evaluated from the right parasternal long-axis view, the right parasternal 4-chamber view, and left apical 4-chamber view. The left atrial-to-aortic root ratio (LA/Ao) was obtained from the right parasternal short-axis view by B-mode method. M-mode measurements of the LV were obtained from the right parasternal short-axis view of the LV at the level of the chordae tendinae, according to the leading edge-to-leading edge method. These included end-diastolic interventricular septum thickness (IVSd), end-diastolic LV free-wall thickness (LVWd), end-diastolic left ventricular internal dimension (LVIDd), end-systolic left ventricular internal dimension (LVIDs), and fractional shortening (FS). LV wall thickness was defined as the sum of the IVSd and the LVWd. As indicators of LV geometrical remodeling, relative wall thickness (RWT) and sphericity index were calculated as previously described. RWT is the ratio of the LV wall thickness to LVIDd, and sphericity index is the ratio of LV long-axis diameter to short-axis diameter in end-diastole. Transmitral flow was obtained from the left apical 4-chamber view, and measured from the early diastolic wave (E wave) and late diastolic wave (A wave). Forward SV was calculated from the product of the aortic valve cross-sectional area (using circular assumption of aortic valve annulus) and the aortic valve outflow time-velocity integral according to the American Society of Echocardiography method. Forward SV was indexed to body surface area to determine the forward SV index (FSVI). Body surface area was derived from body weight using a previously published equation. In our analyses, we used the mean value of 3 consecutive cardiac cycles in sinus rhythm for each measurement.
After completion of the standard conventional 2D, M-mode, and Doppler examinations, high quality images for 2D-STE analysis were carefully obtained by the same investigator and same protocol. For analysis of 2D-STE results, all views were recorded at the rates of 76.1–139.2 frames/s. To evaluate torsional deformations, proper basal and apical short-axis views were carefully obtained using the following anatomic landmarks: at the basal level, mitral valve and at the apical level, LV cavity alone with no papillary muscle visible (Fig 1). Images were analyzed by 1 of 4 trained observers using off-line workstation.3 All observers were blinded to the classification of the dogs under evaluation. We used previously published procedures for speckle-tracking analysis.[14, 22] We selected 1 cardiac cycle in sinus rhythm (from 1 QRS complex to the next QRS complex) from high-quality images and manually traced the endocardial borders of the myocardium to select the appropriate region of interest. The region of interest then was adjusted to incorporate the entire myocardial thickness and checked by trained observers to ensure that it visually synchronized with cardiac movement throughout the entire cardiac cycle. The computer software automatically traced the myocardium and created 6 segments in each image. The software also automatically evaluated whether it reliably followed myocardial motions. If the initial evaluation failed because the region of interest could not be traced during myocardial movement, we retraced the endocardial borders and made corrections as needed. If multiple attempts at evaluation failed, the failed segment was excluded from the analysis. Finally, we stored the values of each frame in each basal and apical view, and obtained peak systolic values of rotation and the rotation rate of each view. Then, we calculated torsion and torsion rate as follows: a net difference of the values of each frame between the basal and apical views, as previously reported in humans[8, 12, 13] and dogs,[14, 22] and obtained peak systolic values of torsion and the torsion rate. Counterclockwise rotation/torsion when viewed from the apex was expressed as a positive value. The mean values of measurements from 3 consecutive cardiac cycles in sinus rhythm from high-quality images were used in all analyses.
Figure 1. Examples of LV basal (A) and apical (B) views for two-dimensional speckle-tracking echocardiography in a MMVD dog. The proper basal level was defined as the view showing the mitral valve leaflet. The proper apical level was defined as the view of LV cavity alone with no papillary muscles visible. Both LV short-axis views were obtained as circular as possible. Six segments were designated as cranial septum (yellow), cranial (light blue), lateral (green), caudal (purple), inferior (dark blue), and septal (red) for speckle-tracking analysis.
Download figure to PowerPoint
Data are expressed as medians and interquartile ranges. All statistical analyses were performed by commercially available computer software.4 We used a Kruskal–Wallis test for testing equality of medians among the 4 groups (controls, I, II, III). Parameters that indicated significant differences by Kruskal–Wallis test among groups then were subjected to pairwise comparison by Mann–Whitney U-tests with Bonferroni's adjustment. A value of P < .008 was regarded as significant. We assessed 2D-STE interobserver variability by determining the coefficients of variation for all measurements of 2 groups of 5 randomly selected dogs. Intraobserver reproducibility was assessed by repeating 2D-STE measurements for 5 dogs selected at random on a different day.
- Top of page
- Materials and Methods
Torsional deformations assessed by 2D-STE differed among clinical classes of MMVD. These deformations may reflect myocardial function. As torsion in compensated dogs with class II MMVD was greater than in asymptomatic dogs with class I MMVD, this may be useful for evaluating compensatory myocardial function and early myocardial dysfunction. In addition, lower peak systolic torsion, possibly suggestive of latent systolic dysfunction, and found in decompensated dogs with class III MMVD compared with controls may contribute to the severe cardiac clinical signs experienced by these dogs. Increasing evidence suggests that assessment of LV function by 2D-STE in humans with chronic severe MR can assist clinicians in determining optimal timing for surgery.[23, 24] Therefore, assessment of definite systolic myocardial function in dogs with MMVD might be clinically useful, especially if surgical mitral valve repair becomes more common. This study is the first to noninvasively measure myocardial torsional deformations and assess myocardial function by 2D-STE in dogs that clinically diagnosed with MMVD.
Given that LA/Ao and LVIDd gradually increased with the progression of the ISACHC classification, volume overload because of MR was higher in more severe classes. This overload resulted in enlargements of the LA and LV. Dogs in class I and II compensate against volume overload, thus avoiding severe cardiac clinical signs. However, the higher heart rate, E wave velocity, E/A ratio, and lower FSVI in class III seem to be related to decompensation that leads to severe congestive heart failure, which is in agreement with previous studies.[5, 25] Volume overload of the LV in dogs with MMVD results in excessive motility as reflected by FS in this study. Increased ventricular wall motion is typical in dogs with advanced mitral valve disease, yet experimental studies demonstrate that the contractility of individual myocardial fibers is decreased.[26, 27] FS provides a limited representation of LV systolic function because of load dependency in dogs with MR. Therefore, we found no definitive indications of systolic dysfunction by conventional echocardiography in the dogs included in this study.
Twisting of the LV is determined by the oblique orientation of the myocardial fibers. The dynamic interaction between subendomyocardial and subepimyocardial fiber helices in the LV creates a twisting motion. The resulting LV torsion is reported to reflect the balance of function and difference in arm length between the subendomyocardium and subepimyocardium. Arm length refers to the length of the subendomyocardial and subepimyocardial radii. In this study, peak systolic torsion, torsion rate, apical rotation, and apical rotation rate were significantly lower in class I than in controls. Because any parameter other than LA/Ao was not significantly different between class I dogs and controls, a comprehensive explanation of the changes in torsional parameters is not obvious, but likely many factors are involved. LV torsion is directly related to the helical orientation of myocardial fibers, and lower torsional parameters may reflect early changes that decrease LV myocardium wall stress. The reduction in wall stress is caused by the reduction in afterload caused by regurgitation.
Difference in arm length between the subendomyocardium and subepimyocardium was lower in class II compared with class I, as indicated by the lower LV wall thickness. Despite this decreased difference of arm length, peak systolic torsion in class II was significantly higher than that in class I. Therefore, the greater torsional deformation observed in class II compared with class I may be caused by a decrease in the myocardial function of the subendocardium and an increase in subepicardium functioning. Gradual remodeling of the LV chamber with the development of thin, elongated myocytes and eccentric hypertrophy is adapted to the increased work required to maintain cardiac output. RWT and the sphericity index (which we used as indicators of LV geometrical differences) decreased with increasing MMVD severity. These findings may indicate that wall stress is increased and that abnormal distribution of myocardial fibers developed in dogs with MMVD. These changes probably restrict subendomyocardial function. In addition, because the subendomyocardial fibers are more sensitive to microvascular ischemia and fibrosis, myocardial function may decrease earlier in the subendocardium than in the subepicardium. Furthermore, because both the subepimyocardium and subendomyocardium work together to produce a nearly homogenous transmural distribution of myocardial stress and fiber strain during ejection, it may be possible to compensate for decreases in subendomyocardial function by increasing subepimyocardial function. Therefore, decreased myocardial function in the subendocardium combined with compensatory increases in subepicardium function may cause the torsional deformation observed in class II. Clinically, high torsional deformation may be useful for evaluating compensatory myocardial function and would be a sensitive indicator of myocardial dysfunction in dogs with MMVD.
Dogs with class III MMVD have severe cardiac clinical signs and peak systolic torsion was lower compared with that of compensated dogs with class II MMVD and controls in this study. Previous studies have demonstrated that systolic torsion was decreased in animal models of chronic MR compared with the acute phase in the same individuals.[31, 32] Torsional deformation in class III dogs may result from limited compensatory myocardial function in the subepicardium and a reduction in the arm length of the subendomyocardium and subepimyocardium, as indicated by the thinner LV walls observed in class III compared with controls. The amount of wall stress may be higher and the distribution of myocardium may be more abnormal in class III than in other MMVD classes and controls. This hypothesis is supported by the lower RWT and sphericity index of class III dogs compared with class II dogs and controls. Increases in wall stress and abnormal myocardium may restrict subendocardial and subepicardial myocardial functions. Cellular remodeling[26, 27] and sarcomere dysfunction in the myocardium may contribute to fiber shortening as chronic MR progresses. Thus, twisting motions generated by the myofibers may decrease in advanced classes of MMVD, leading to the small torsion we observed in class III. Although some of the dogs in class II and III were receiving medication at the time of this study, many torsional parameters of class III dogs were not higher than those of dogs in other classes, and peak systolic torsion was lower in class III than in class II and controls. This observation also suggests myocardial dysfunction in dogs with severe MMVD.
Torsion is thought to be a mechanism by which the ventricle equalizes transmural oxygen demand gradients, thus minimizing oxygen consumption and optimizing myocardial energetics and efficiency.[34-36] Impaired myocardial functions are associated with the progression of clinical signs and poor outcomes for both humans and dogs with MR.[27, 37] Torsional deformation may be related to the cardiac clinical signs of dogs with MMVD. The evaluation of torsional deformation therefore has the potential to allow early identification of dogs at risk for progression to severe MMVD.
This study has several limitations. First, when calculating torsion from basal and apical rotation, the number of frames showing basal and apical views should correspond exactly. However, there was no way to obtain views with exactly the same number of frames. We tried to correct these views carefully, to ensure that they represented conditions that were as similar as possible. Second, because our study was a noninvasive clinical study, we could not measure myocardial contractility definitively by the invasive assessment that is the gold standard for assessing LV function. Third, assessment of pulmonary hypertension in this study was dependent on identification of tricuspid regurgitation detected by Doppler echocardiography. It is therefore possible that some dogs with pulmonary hypertension and no evidence of tricuspid regurgitation were included in the analysis. Fourth, we did not assess myocardial deformation by 2D-STE in the same dogs. We were therefore unable to relate changes in myocardial deformation to deterioration or improvement of MMVD. Finally, we did not consider the influences of medications on the measured parameters. These limitations should be addressed in future investigations.
In conclusion, torsional deformations assessed by 2D-STE differed among clinical classes of MMVD. The lower torsion in dogs with severe MMVD may contribute to latent systolic dysfunction and seems to be related to severe cardiac clinical signs. Myocardial torsional deformations by 2D-STE may provide more detailed assessment of contractile function in dogs with MMVD. Nevertheless, the clinical importance of myocardial torsional deformations by 2D-STE and their relationship with survival time needs to be investigated.