Mechanism of Beneficial Effects of Restrictive Mitral Annuloplasty in Patients With Dilated Cardiomyopathy and Functional Mitral Regurgitation
Background—Restrictive mitral annuloplasty (RMA) often leads to reverse left ventricular (LV) remodeling in patients with advanced cardiomyopathy and functional mitral regurgitation. However, the mechanism responsible for its favorable effects on LV ejection performance has been poorly understood. We evaluated systolic wall stress using cineangiographic multidetector computed tomography (MDCT) and our developed software system to assess stress-shortening relations before and after RMA.
Methods and Results—Twenty-four patients with dilated cardiomyopathy underwent 64-row MDCT before and 2 months after RMA. All patients underwent stringent downsizing annuloplasty with a semirigid complete ring. Reconstructed images were used to calculate LV end-diastolic index (EDVI) and end-systolic volume index (ESVI), LV ejection fraction, and regional and global end-systolic wall stress (ESS). After RMA, LVEDVI and LVESVI decreased from 151±52 to 131±53 mL/m2 (P=0.0001) and from 114±48 to 92±50 mL/m2 (P=0.0001), respectively. Global ESS decreased from 157±43 to 139±50 kdyne/cm2 (P=0.01), and LV ejection fraction improved from 27±8.0 to 33±13% (P=0.0007). There were significant correlations between change in LVEDVI and LVESVI (r=0.88, P<0.0001) and change in LVESVI and global ESS (r=0.68, P=0.0002). Moreover, the magnitude of increase in LV ejection fraction significantly correlated with the degree of reduction in global ESS (r=−0.61, P=0.002). Patients without significant reverse LV remodeling had significantly higher preoperative and postoperative global ESS than those with it.
Conclusions—Our study suggests that decrease in afterload after reduction in volume overload was responsible for postoperative reverse LV remodeling process after RMA.
Functional mitral regurgitation (MR) is a common complication in both ischemic and nonischemic dilated cardiomyopathy and leads to an increased left ventricular (LV) volume overload. This condition further aggravates LV remodeling process, eventually leading to congestive heart failure.1 To resolve this vicious cycle, Bolling et al2 first reported the feasibility of undersized mitral annuloplasty in patients with end-stage cardiomyopathy. They observed reverse LV remodeling (decrease in LV volumes and increase in ejection fraction) after correction of MR in such failing ventricle, although further decline in LV ejection performance was to be expected as a result of increase in afterload (systolic myocardial wall stress) accompanied by the abolition of the low-impedance pathway for ejection into the left atrium.3 Thereafter, several studies demonstrated that undersized mitral annuloplasty or restrictive mitral annuloplasty (RMA) procedure could induce reverse LV remodeling.4–8
However, to date, the mechanism how RMA can lead to reverse LV remodeling process remains unclear. In the present study, we hypothesized that change in systolic myocardial wall stress may be a key explaining of reverse LV remodeling after correction of functional MR in dilated cardiomyopathy. We recently developed a software package to determine local myocardial stress using cineangiographic 64-row multidetector computed tomographic (cine-MDCT) images.9 To test our hypothesis, therefore, we used both cine-MDCT images and our developed software.
The study group consisted of 24 patients (19 men and 5 women, mean age of 64±9.4) who underwent 64-row MDCT before and after RMA between January 2007 and December 2008. All patients had a diagnosis of advanced dilated cardiomyopathy with LV dysfunction (LV ejection fraction <40%) and congestive heart failure symptoms despite the maximal medical treatment. All patients had 3 to 4+ functional MR secondary to LV and annular dilatation and systolic restrictive motion of mitral leaflets on echocardiography. Eleven patients also had coronary artery disease and were classified as having ischemic dilated cardiomyopathy.10 The study protocol was approved by the institutional review board, and all patients provided informed consent. Table 1 shows preoperative clinical profile of all patients.
All operation was performed with conventional cardiopulmonary bypass with mild hypothermia. Myocardial protection was achieved by both antegrade and retrograde cold blood cardioplegia. All patients underwent a stringent RMA with the implantation of semirigid complete ring (Carpentier Edwards Physio Ring; Edward Lifesciences, Irvine, Calif). The size of annuloplasty ring used was 24 mm in 12 (50%), 26 mm in 11 (46%), and 28 mm in 1 (4%). Simultaneous procedure included coronary artery bypass grafting in 11 (46%) and tricuspid annuloplasty in 15 (63%).
MDCT Image Acquisition
Image acquisition was performed 1 month before surgery and repeated at 2.3 months after surgery. During each image acquisition, upper extremity blood pressure in substitution for invasive measurement of LV end-systolic pressure was recorded with an automated digital cuff. After an initial scout image and a timing bolus scan, a volume data set of the whole heart (collimation, 64×0.625 mm; gantry rotation time, 350 ms; mean scanning time, approximately 5 seconds) was acquired with retrospective electrocardiographic gating. For enhancement of LV cavity, 0.8 mL/kg of nonionic contrast medium was administered at a flow rate of 4 mL/s. Ten data sets were obtained at 10% R-R interval increments throughout the cardiac cycle (5% to 95%). A stack of consecutive images (each of 8 mm thick and a gap of 2 mm) was reconstructed in the short-axis of LV. To calculate local end-systolic wall stress (ESS), LV horizontal and vertical long-axis cine-images according to standardized LV segmentation11 were reconstructed throughout one cardiac cycle and stored as audio video interleave files using commercially available software (Aquarius Net Station, Terarecon, Inc, San Mateo, Calif). Image acquisitions were successful without any complications in all patients. All image processing were verified by a well-experienced radiologist (S.H.).
MDCT Image Analysis
Each long-axis cine-image was transferred to off-line personal computer and further analyses were performed with the use of our developed software (YD, Ltd, Ikoma, Nara, Japan). Details of logic and validity of the software have been published previously.9 The image with the smallest LV chamber area was selected as the end-systolic one, and the image with the largest LV chamber was defined as the end-diastolic one. LV measurements including LVEF, mass, LV end-diastolic volume (LVEDV) and end-systolic volume (LVESV) were performed according to the slice summation method by manually drawing the endocardial and epicardial contours of short-axis images. Papillary muscles were included in the LV cavity. LV volumes were indexed by the body surface area (LVEDVI and LVESVI, respectively). Local ESS was calculated on the basis of the Janz equation 112; ESS=P×ΔAc/ΔAw, where P is LV end-systolic pressure, and ΔAc and ΔAw are local cross-sectional area of the LV cavity and local cross-sectional area of the LV wall at end-systole, respectively. In this study, LV end-systolic pressure estimate was obtained with a following equation: P=0.98×(systolic blood pressure+2×diastolic blood pressure)/3+11 mm Hg. The rational for using this equation in the setting of chronic MR was described previously.13 The software can measure local ESS along 100 chords evenly spaced along the epicardial border in each image (Figure 1). In the present study, to simplify the display and analysis of local ESS, 3 LV levels (base, mid, and apex) were determined with reference to long axis and 3 perpendicular short axes at equal intervals and each level was subdivided into 4 segments (anterior, inferior, lateral, and septum). Regional ESS was calculated by averaging the value at consecutive local chords within each level and segment. To examine the relationship between changes in global LV function and ESS, global ESS was defined as the average value at all consecutive chords obtained. Finally, to describe the global LV shape, the sphericity index was calculated as the LV minor-axis diameter divided by the LV major-axis diameter on the end-diastolic and end-systolic horizontal long-axis image. All measurements were repeated 2 times, and the average value was used as a final one.
All patients were assessed by 2-dimentional echocardiography around the same time as MDCT study. Degree of mitral regurgitation was quantified by means of echocardiographic Doppler analysis using a 5-point scale (from 0 to 4+).
The SPSS (version 11.0, SPSS Inc) software was used for statistical analysis. The continuous variables are presented as mean±SD and compared using paired and unpaired t test. Categorical variables were compared by χ2 analysis and Fisher exact test. Correlation between variables was tested with Pearson correlation analysis. Statistical significance was defined as P<0.05.
Surgical and Late Clinical Results
All patients well tolerated the operation. On echocardiography, the degree of MR was less than grade 2+ in all patients after surgery and significantly decreased from 3.3±0.6 to 0.3±0.5 (P<0.0001). During a mean follow-up of 18 months, 5 patients (21%) had recurrent congestive heart failure symptoms in NYHA (New York Heart Association) class III or IV. The other patients were well in NYHA class I or II.
Changes in LV Ejection Performance and Wall Stress Before and After RMA
Heart rate and blood pressure at the time of image acquisition did not significantly change postoperatively (Table 2). LVEDVI significantly decreased with an average reduction of 20 mL/m2 (13% reduction) and LVESVI also decreased with an average reduction of 21 mL/m2 (21% reduction) (Table 3). Mean LVEF significantly improved from 27% to 33% after surgery. The measured value of postoperative LVEF was higher than the predictive value (28%) when we assumed that stroke volume did not change after surgery. Reduction percentage in LVEDVI and LVESVI and increase in LVEF were larger in patients with ischemic etiology compared with those with nonischemic etiology; ΔLVEDVI (−16% versus −9.4%, P=0.27), ΔLVESVI (−27% versus −13%, P=0.16), and ΔLVEF (+9.6% versus +3.2%, P=0.07); however, those differences did not reach statistical significance as of 2-month follow-up.
LV mass index did not significantly change. Both end-diastolic and end-systolic sphericity indices significantly decreased after the operation. The ratio of global ESS to LVESVI (ESS/ESVI), which was a relatively load-independent index for estimating myocardial contractility,14,15 significantly improved from 1.5±0.5 to 1.7±0.5 (P=0.008), with a mean increase of 17% (Figure 2).
Change in LV Global and Regional Wall Stress After RMA
Change in regional ESS at 12 segments was shown in Figure 3. Regional ESS was generally reduced at all segments, especially at anterior and inferior segments. Consequently, global ESS significantly decreased with a mean reduction of 11% (Table 3). Postoperative reduction in average ESS was larger in midventricular level (13% reduction) than in basal and apical ventricular level (9% reduction for each).
Relationship Between LV Volume, Wall Stress, and Ejection Performance
There was a strong positive correlation (r=0.88, P<0.0001) between ΔLVEDVI and ΔLVESVI (Figure 4A). ΔLVESVI was significantly related to Δglobal ESS (r=0.68, P=0.0002) (Figure 4B). ΔGlobal ESS significantly correlated with ΔLVEF (r=−0.61, P=0.002) (Figure 4C). Moreover, weak correlation was noted between ΔESS/ESVI and ΔLVEF (r=0.48, P=0.02). ΔEnd-systolic sphericity index weakly but significantly correlated with Δglobal ESS (r=0.46, P=0.02).
Impact of ESS Reduction on Postoperative Clinical Outcomes
When significant reverse remodeling was defined as a volume reduction exceeding 15% in LVESVI,6 17 patients (71%) showed significant reverse remodeling and 7 (2 in ischemic and 5 in nonischemic) showed no or poor reverse remodeling 2 months after RMA. Patients without significant reverse remodeling had higher preoperative global ESS (186±20 versus 143±44 kdyne/cm2, P=0.01) and persistently higher postoperative global ESS (190±35 versus 119±41 kdyne/cm2, P=0.0004) than those with it (Figure 5), whereas preoperative degree of LV remodeling was similar between groups (LVESVI; 127±40 versus 107±51 mL/m2, P=0.34). Four of 7 patients (57%) without significant reverse LV remodeling developed late recurrent heart failure, compared with 1 of 17 patients (5.9%) with it (P=0.02).
Major findings of this study were (1) ESS significantly decreased after RMA, and the reduction in ESS was strongly associated with the improvement in LVEF; and (2) ESS/ESVI significantly improved, and the improvement in ESS/ESVI was weakly but significantly correlated with the improvement in LVEF. These findings indicate that the improvement in LV ejection performance was mostly related to the afterload reduction rather than to the improvement of LV contractility as of 2 months after the operation. That is, the mechanism of reverse LV remodeling process is related to the afterload reduction derived from ventricular volume unloading by eliminating MR.
It was once believed that the presence of chronic MR creates a systolic unloading effect by providing a low-resistance ejection into the left atrium. This traditional hypothesis held that the mitral valve functions as a “pop off valve” for the failing ventricle and surgical correction might have prohibitive risks in patients with severely remodeled LV because of postoperative afterload excess and subsequent decline in LV ejection performance.3 However, several recent studies have shown an acceptably low operative mortality and have suggested that RMA can lead to reverse LV remodeling.2,4–8 Consequently, it has been postulated that RMA could reduce myocardial wall stress and promote reverse LV remodeling2; however, change in regional myocardial stress after RMA has not been fully investigated. To our knowledge, this is the first study to assess the changes in systolic wall stress before and after RMA in patients with cardiomyopathy using cine-MDCT images.
Our software demonstrated that regional ESS largely decreased after RMA, leading to global ESS reduction. Few studies have examined the change in wall stress after valve repair in patients with advanced cardiomyopathy. A Previous experimental report demonstrated that annular reduction could reduce regional ESS by reducing radii of curvatures at basal, equatorial, and even the apical levels without any reduction in LV volume.16 These data appeared to support Bolling’s concept that undersizing the annulus acutely remodels the base of the heart, giving the ventricle a more elliptical shape (a direct reshaping effect) and thereby decreased LV wall stress.2 However, in the present study, reduction in ESS was mostly related to LV volume reduction. Both stronger correlation between ΔLVESVI and Δglobal ESS (r=0.68) compared with that between Δsphericity index and Δglobal ESS (r=0.46) and lesser ESS reduction in basal LV (9%) compared with that in mid LV (13%) might suggest that the effect of volume unloading had a much greater impact on ESS reduction than the direct reshaping effect. Therefore, LV unloading and subsequent reduction in wall stress would lead to reversal of ventricular remodeling and favorable change in LV shape. The issue about the relative impact of direct annular remodeling on change in ESS requires further investigation.
There is accumulating evidence that reverse LV remodeling has been demonstrated as a result of either annuloplasty alone or in combination with coronary artery bypass grafting.2,4–8 Bolling et al2 observed 8% reduction in EDV and 15% reduction in ESV with the improvement in LVEF from 0.18 to 0.27 at 6 months after the operation. Subsequently, they also reported 23% reduction in EDV and 28% reduction in ESV with the change in LVEF, from 0.20 to 0.26 at a later period.4 Acker et al8 also reported that significant reverse LV remodeling persisted up to 18 months after mitral valve surgery in patients with nonischemic heart failure. These findings suggest that reverse LV remodeling is a gradual and time-dependent process. Many other echocardiographic studies also have reported that significant decrease in LV dimensions and increase in LVEF (“reverse remodeling”) in most of patients. Our data were largely consistent with those of previous reports; 71% of patients showed significant reverse remodeling at 2-month postoperative MDCT study. However, persistent or progressive remodeling pattern is also documented despite an initially successful mitral repair with longer follow-up. De Bonis et al7 reported that of 111 patients with advanced cardiomyopathy who underwent RMA, 52% of their patients showed reverse LV remodeling, whereas 48% showed unchanged or progressive remodeling at 2-year follow-up. Gelsomino et al6 observed significant reverse LV remodeling in 44% of 251 patients undergoing RMA for ischemic MR and continued or progressive LV remodeling in 56% at 5 years after the operation. These conflicting results are likely to be attributed to the complex mechanisms of functional MR itself, which include not only annular dilatation but also papillary muscle displacement and chordal tethering caused by LV remodeling.17 Although RMA could improve mitral leaflet coaptation,18 most recent studies demonstrated that simple annuloplasty repair could not necessarily improve leaflet tethering enough to avoid recurrent MR19 and therefore yielded the absence of reverse LV remodeling.6,7 Current results demonstrated that the resolution of annular problem by RMA could provide, at least, less of a stimulus for ongoing pathological myocardial remodeling process but could not restore normal LV function and configuration as seen in the nonfailing hearts.20 The complex mechanism involving the valvular and subvalvular apparatus might require more comprehensive reparative correction rather than simple annular repair to improve the results.8
Our result demonstrated that reverse LV remodeling process after RMA was induced by reduction of not only volume overload (ventricular preload) but also regional and global systolic wall stress (afterload). Change in ejection performance after RMA results from a complex interaction of contractile function, ventricular preload, and afterload. Reduced ventricular preload indicated by the reduction in EDV (eliminating MR) would have actually tended to reduce ejection performance.3 However, our data indicated that reduced ventricular preload also allowed reduced systolic chamber size and elliptical LV shape, which resulted in reduction of the radius term in the wall stress equation, allowing systolic wall stress to be reduced. Reduced afterload allowed further reduction in ESV, overcoming the fall in EDV to allow ejection performance to improve. Furthermore, our result demonstrated modest increase in ESS/ESVI, suggesting that the potential benefits of RMA might extend to an improvement in myocardial contractility. The relationship between ΔLVEF and ΔESS/ESVI indicated that myocardial contractile recovery might contribute a favorable effect to reverse remodeling process. However, the implication of this improvement should not be overestimated. Carabello et al14 reported that an ESS/ESVI of <2.6 indicated a poor prognosis in patients with MR. Therefore, this positive effect might not be clinically relevant because postoperative mean ESS/ESVI of 1.7 remained extremely low in our cohort. Indeed, some previous studies demonstrated that the role of valve repair for the failing hearts was limited in terms of survival benefits. Enriquez-Sarano et al21 showed the ongoing poor effects of mitral repair on 15-year survival in patients with ischemic MR. Braun et al5 reported the limited effect of RMA in patients with LV end-diastolic dimension >65 mm. These results imply that beneficial effects of RMA on LV performance might not be enough to overcome the intrinsic LV failure. To date, there has been no conclusive data about the true effect of mitral valve surgery on the prognosis. Although 2 large series concluded no clear survival benefit of annuloplasty repair for functional MR compared with medical therapy or isolated coronary bypass grafting,22,23 those were limited because of retrospective study design. It is hoped that randomized, prospective trials will resolve this problem.
Our result partly explains the failure of reverse remodeling process in terms of measured LV wall stress using our software. Grossman et al24 hypothesized that compensatory mechanism to maintain physiological myocardial wall stress (afterload) could be mediated through the addition of new series sarcomeres and the development of eccentric myocardial hypertrophy in chronic volume-overload LV. However, inability to sustain addition of new sarcomeres in progressive volume overload could result in high myocardial wall stress, leading to decreased myocardial fiber shortening and ventricular failure (decompensated MR). Our data suggest that RMA could partly attenuate this pathological myocardial process even in the decompensated MR. However, the fact that patients with preoperative high global ESS showed persistence of the high ESS, which was probably a major factor behind the reduced postoperative ejection performance and, therefore, the adverse clinical outcomes, implied the onset of irreversible pathological process of myocardial dysfunction.25 Wall stress is a major determinant of LV architecture, function, and myocardial composition. Therefore, further analyses using our software might become helpful to assess the reverse LV remodeling viability in patients with decompensated MR.
There are several limitations in this study. The patient cohort is small and relatively heterogeneous. Moreover, various surgical procedures added and type of annuloplasty ring used may influence the outcomes. In our series, change in LVESVI and LVEF was not significantly different between ischemic and nonischemic group as of 2 months after the operation. Moreover, we have reported that revascularization alone could not always decrease regional stress in patients with LV dysfunction.26 Therefore, we think that the effect of revascularization could be minimized in our results. However, those differences might become obvious with further follow-up, because the effect of revascularization can be considerably delayed over 3 months in the failing ventricle.27 We preferably used semirigid complete annuloplasty ring. Previous report suggested that the type of annuloplasty ring could influence the postoperative LV function28; therefore, our results would not be applicable to patients who received another type of prosthesis.
We estimated local ESS using the simplified formula derived by Janz12 because this method is easy to apply when clinically determining local ESS. Although the methodological problems have been discussed in previous reports,9,26 it is considered that values obtained by the Janz method are comparable to those obtained by the finite element method within an error of ≤10%.
Finally, we used a 64-row MDCT as the image modality. As found in other conventional modalities, there are inherent limitations, including the need for contrast medium and radiation exposure. However, to estimate local wall stress, its high spatial resolution enables us to precisely estimate regional wall thickness and appropriately compare preoperative values with postoperative values. Moreover, because of current increasing use of cardiac defibrillator or biventricular pacemaker in patients with heart failure, we think that MDCT is useful to noninvasively assess accurate ventricular function.29,30
In conclusion, RMA could reduce regional and global wall stress by eliminating ventricular preload. Our study suggested that the extent of reduction in afterload was partly responsible for postoperative reverse LV remodeling process after RMA.
Presented at the 2009 American Heart Association meeting in Orlando, Fla, November 14–18, 2009.
Mann DL. Mechanisms and models in heart failure: a combinatorial approach. Circulation. 1999; 109: 999–1008.
Schuler G, Peterson KL, Johnson A, Francis G, Dennish G, Utley J, Daily PO, Ashburn W, Ross J Jr. Temporal response of left ventricular performance to mitral valve surgery. Circulation. 1979; 59: 1218–1231.
Gelsomino S, Lorusso R, De Cicco G, Capecchi I, Rostagno C, Caciolli S, Romagnoli S, Da Broi U, Stefàno P, Gensini GF. Five-year echocardiographic results of combined undersized mitral ring annuloplasty and coronary artery bypass grafting for chronic ischaemic mitral regurgitation. Eur Heart J. 2008; 29: 231–240.
De Bonis M, Lapenna E, Verzini A, La Canna G, Grimaldi A, Torracca L, Maisano F, Alfieri O. Recurrence of mitral regurgitation parallels the absence of left ventricular reverse remodeling after mitral repair in advanced dilated cardiomyopathy. Ann Thorac Surg. 2008; 85: 932–939.
Shudo Y, Matsumiya G, Takeda K, Matsue H, Taniguchi K, Sawa Y. Novel software package for quantifying local circumferential myocardial stress. Int J Cardiol. 2010 [Epub ahead of print].
Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S, Laskey WK, Pennel DJ, Rumberger JA, Ryan T, Verani MS. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the cardiac imaging committee of the council on clinical cardiology of the American Heart Association. Circulation. 2002; 105: 539–542.
Rozich JD, Carabello BA, Usher BW, Kratz JM, Bell AE, Zile MR. Mitral valve replacement with and without chordal preservation in patients with chronic mitral regurgitation: mechanism for differences in postoperative ejection performance. Circulation. 1992; 86: 1718–1726.
Carabello BA, Williams H, Gash AK, Kent R, Belber D, Maurer A, Siegel J, Blasius K, Spann JF. Hemodynamic predictors of outcome in patients undergoing valve replacement. Circulation. 1986; 74: 1309–1316.
Taniguchi K, Nakano S, Kawashima Y, Sakai K, Kawamoto T, Sakaki S, Kobayashi J, Morimoto S, Matsuda H. Left ventricular ejection performance, wall stress, and contractile state in aortic regurgitation before and after aortic valve replacement. Circulation. 1990; 82: 798–807.
Levine RA, Schwammenthal E. Ischemic mitral regurgitation on the threshold of a solution: from paradoxes to unifying concepts. Circulation. 2005; 112: 745–758.
Shudo Y, Matsue H, Toda K, Hata H, Fujita S, Taniguchi K, Sawa Y. A simplified echocardiographic measurements of direct effects of restrictive annuloplasty on mitral valve geometry. Echocardiography. 2010 (in press).
Lee AP, Acker M, Kubo SH, Bolling SF, Park SW, Bruce CJ, Oh JK. Mechanism of recurrent functional mitral regurgitation after mitral valve repair in nonischemic dilated cardiomyopathy: importance of distal anterior leaflet tethering. Circulation. 2009; 119: 2606–2614.
Enriquenz-Sarano M, Schaff HV, Frye RL. Mitral regurgitation: what causes the leakage is fundamental to the outcome of valve repair. Circulation. 2003; 108: 253–256.
Bondarenko O, Beek AM, Twisk JW, Visser CA, van Rossum AC. Time course of functional recovery after revascularization of hibernating myocardium: a contrast-enhanced cardiovascular magnetic resonance imaging. Eur Heart J. 2008; 29: 200–205.