Mitral regurgitation in dilated cardiomyopathy: value of both regional left ventricular contractility and dyssynchrony

Abstract

Aims

Mitral regurgitation (MR) is common and independently predicts mortality in patients with left ventricular (LV) systolic dysfunction. Its management remains challenging because of the complexity and variety of potential mechanisms implicated. We sought to determine which LV functional characteristics are the most important determinants of the severity of the MR associated with dilated cardiomyopathies.

Methods and results

We performed echocardiographic studies in 87 consecutive patients with dilated cardiomyopathy. The degree of MR was quantified according to guidelines. LV, left atrial and mitral annulus dimensions, mitral valve tenting, estimated filling pressures, regional myocardial contractility, and dyssynchrony (using regional strain (ε) analysis) were recorded too. Determinants of significant MR was thus assessed using multivariate models. Mitral regurgitant volume correlated with mitral annulus diameter ( P < 0.001), mitral valve tenting height ( P < 0.001), LV volumes ( P = 0.004), LV ejection fraction, mid-lateral wall peak of ε ( P = 0.01), and its delay ( P < 0.001). That inter-relation between the mitral annulus, the ventricle shape, contractility, and dyssynchrony was founded in the multivariate analysis. As a matter of fact, the model predicting the best the MR volume ( R = 0.78) included: mitral annulus diameter, dyssynchrony, tenting heigh and contractility of the LV mid-lateral wall (measured by ε).

Conclusion

The MR of the dilated cardiomyopathy is multifactorial. Our data suggest that analysing only LV geometry and mitral orifice is insufficient to correctly describe functional MR determinant. LV contractility and dyssynchrony are essential too.

Introduction

Chronic heart failure (CHF) due to systolic dysfunction is a leading cause of cardiac morbidity and mortality. ^{1 , 2} This evolving clinical syndrome is the final common pathway of a large variety of disease entities. ³ Progressive left ventricular (LV) remodelling leading to CHF is often associated with functional regurgitation of a structurally normal mitral valve, ^{4 , 5} which predicts poor haemodynamic and clinical outcomes. ^{6 , 7} Our understanding of the anatomic and functional determinants of the severity of mitral regurgitation (MR) in CHF remains limited, ^8–11 and its management remains challenging. ^12–14 It depends on surgical systems and procedures, such as cardiac support devices or restoration of the LV geometry by repairing the ventricle or mitral valve annuloplasty. ^{15 , 16} A few recent studies have reported beneficial effects conferred by cardiac resynchronization therapy (CRT). ^12–14 In addition, the CARE-HF echocardiographic substudy has shown favourable, though modest, effects of CRT on the degree of MR. ^{8 , 17}

Tissue Doppler imaging has been proposed to study LV mechanical dyssynchrony and LV regional contractility. ^{18 , 19} Strain (ε,%) modality measures regional myocardial deformation, and thus distinguishes passive displacement from active systolic contraction, in opposition to the assessment of myocardial velocities. ¹⁹ This prospective, single-centre study of consecutive patients referred for echocardiographic evaluation of CHF was performed to improve comprehension of the complex mechanisms implicated in functional MR. In particular, we sought to examine the relationships between severity of MR and ventricular, atrial and mitral annulus dimensions, and intra-LV dyssynchrony and regional deformation.

Methods

Study population

We prospectively enrolled consecutive patients in stable CHF (NYHA class II to IV) treated with recommended drug treatments. Every patient had an LV ejection fraction <40% and echocardiographic end-diastolic diameter ≥55 mm. It was requested to have more than a trivial MR to be included in the study (≥2/4).

Patients with organic mitral valve disease and patients with myocardial infarction were excluded considering their mechanism of MR distinct from the one of dilated cardiamyopathy (DCM)-population as demonstrated by Agricola et al . ²⁰ This study was reviewed and approved by our Ethics Committee, and all patients agreed with the study.

Echocardiographic study protocol

Transthoracic echocardiography was performed using a ViVid Seven (GE Healthcare, Milwaukee, WI). Five-beat loops of apical four-chamber views in 2-D grey scale and colour Doppler tissue imaging were stored. The image sector was ∼30°, as narrow as possible to reach a rate of >140 frames/s in the Doppler tissue-imaging mode.

Quantification of MR

A careful re-analysis of the images case by case using every approach available allowed a quantification of the degree of MR. ^{21 , 22} The mitral valve regurgitant volume was calculated from the PISA-method in mid-systole and using the continuity equation one. The concordance between the two approaches was requested. If a discordance was observed, the images were reanalysed. At the end, the correlation analysis was performed based on the continuity equation quantification.

Mitral apparatus description

Antero-posterior mitral annular diameter was measured in the parasternal long axis. Mitral valvular tenting height in mid-systole was also measured on the parasternal long axis view, using zoomed loops focused on the mitral valves. ²²

Global LV remodelling

LV end-diastolic and end-systolic volumes (bi-plane Simpson disc method) and diameters were measured. LV short- and long-axis dimensions were measured at the end of diastole and end of systole. Their ratio (sphericity index) was calculated at endsystole and enddiastole. ²³ Efforts were made to record the apical four-chamber view in inspiration to optimize long-axis measurement.

Filling pressure and left atrial dimension assessment

Filling pressures were estimated according to guidelines. ²⁴ We measured both the ratio E/Ea and E/Vp considering Ea as the peak velocity in early diastolic at the mitral annulus (lateral side) (supposed high filling pressure: E/Ea > 15). Vp was the colour M-mode velocity of LV filling in early diastole (supposed high filling pressure: E/Vp > 2.5). LA dimensions considered for the analysis were LA diameter, using the M-mode in parasternal long axis view, and LA area, measured in apical four-chamber view. These measurements of the LA were performed in end-diastole.

Mechanical dyssynchrony assessment

We assessed the presence and degree of both inter- and intraventricular dyssynchrony. Interventricular dyssynchrony was assessed as the time interval between pre-aortic and pre-pulmonary ejection times, as previously described. ²⁵ Intra-ventricular dyssynchrony was assessed as the time delay between the earliest and the latest peak values of longitudinal strain recorded in the mid-segment of the lateral and septal walls. The peak of strain was measured in end-systolic (closure of the aortic valve).

Analyses were performed in mid-segments in order to be close to the papillary muscles ( Figure  1 ). A normalization of time delays was performed. ²⁶

Statistical analysis

Continuous variables are presented as means ± SD. Paired Student’s t -test was performed to examine differences between measurements made in patients with greater than or equal to vs. less significant MR. Bivariate analyses were performed using Pearson’s correlation and Mann–Whitney’s test. Ascendant and descendant linear regression analyses were performed to examine the relationship between severity of MR and valvular and myocardial variables. For these analyses, the normality of the residue hypothesis was tested using Kolmogorov–Smirnov test, and the non-autocorrelation of residue was tested with the Durbin–Watson test. The homoscedasticity hypothesis was also tested, using the SPSS version 10.0 statistical software (SPSS Inc., Chicago, IL). A P -value <0.05 was considered statistically significant.

Intra- and inter-observers variabilities were calculated as percentage difference. The percentage difference was calculated for each pair of measurements as [measurement 1 − measurement 2]/[0.5 × (measurement 1 + measurement 2)] × 100 and expressed as mean ± standard deviation. ²⁷

Results

Feasibility

32% of the images had to be re-analysed to obtain a concordance between the quantification of the MR using the continuity equation and using the PISA method. In functional MR, the PISA method was usually under-estimated the degree of regurgitation.

Reproducibility

The intra-observer variability considering 13 patients re-analysed following the same methodology is as follows.

1. The assessment mitral regurgitant volume using the PISA method was 7.6 ± 5.4%.

2. The measurement of the peak of strain was 12 ± 14%.

3. The time delay between the QRS and the peak of strain in the mid-segment of the lateral wall was 4.9 ± 3.4%.

The inter-observer variability considering 13 patients reanalysed is as follows.

1. The assessment mitral regurgitant volume using the PISA method was 9.7 ± 6.2%.

2. The measurement of the end-systolic peak of strain was 17 ± 10%.

3. The time delay between the QRS and the peak of strain in the mid-segment of the lateral wall was 6.1 ± 3.5%.

Baseline characteristics of patients

The population is described in Table  1 .

We did a comparison of the population according to the presence or the absence of a left bundle branch block (LBBB) with a QRS > 120 ms.

Thirty-five patients had an LBBB. The MR volume was not significantly different from patients with QRS ≤ 120 ms (39.6 ± 12.8 vs.34 ± 0.27 ml, P = 0.3). The LV diameters and volumes were not significantly different. Only the sphericity index measured in diastole tended to be greater in the LBBB patients (1.32 ± 0.25 vs. 1.28 ± 0.28, P = 0.05). Also the mitral tenting height tended to be greater in the LBBB population (0.84 ± 0.27 vs. 0.69 ± 0.29, P = 0.06). The inter-ventricular dyssynchrony was significantly more important in the LBBB group but no significant difference was observed in regard to intra-LV dyssynchrony (473 ± 96 vs. 450 ± 92 ms; P = 0.1) and peaks of deformation (ε(%) in the lateral wall: 9.6 ± 4.6 vs. 8.5 ± 3.8, P = 0.1).

Correlated factors with the degree of MR; univariate analysis

Table  2 displays the univariate correlations between MR Rvol and indices of mitral apparatus and of LV function, size, and dyssynchrony. The correlation and the linear regression analyses between mitral Rvol and a) mitral annulus diameter and b) mitral valve tenting height are shown in Figure  2 , and with a) the degree of LV-myocardial deformation (ε) and b) its time-delay are shown in Figures  1 and 2 .

The analysis as shown in Figure  3 was performed considering the patients having the most severe MR (regurgitant volume > median value of the 80 patients studied, n = 43).

Multivariate analysis

In the multivariate regression analysis, the best predictive model for severity of MR ( R -value = 0.78, P < 0.001), based on the Rvol, was composed of:

Mitral orifice characteristics:

1. mitral annulus diameter,

2. tenting height.

But also, LV characteristics:

1. intra-LV dyssynchrony,

2. peak of maximal deformation in the LV lateral wall.

Correlation between mitral valve tenting height and lv characteristics

The tenting height did not correlate significantly with LV diameters (with end systolic diameter: R = 0.32) or volumes (with end-systolic volume, R = 0.24). The correlation between the tenting height and the sphericity index measured in systole and diastole were not significant too (with end systolic sphericity index, R = −0.10).

The tenting height did correlate with LV dyssynchrony and contractility. The Pearson R -value was 0.36, P = 0.04 between tenting height and the peak of strain recorded in the LV lateral wall. R was 0.41, P = 0.02 between tenting height and the delay of LV lateral wall peak of strain.

Discussion

This study demonstrates that the degree of functional MR is determined not only by mitral orifice characteristics but also by LV characteristics. The most important LV characteristics are not size or geometry but longitudinal contractility (strain of LV mid-lateral wall) and dyssynchrony.

Importance of LV geometry, its consequence on subvalvular apparatus, and mitral annulus shape

Several competing geometric and haemodynamic factors may cause functional MR. ^{9 , 11 , 12 , 14 , 21} The respective importance of tethering of the leaflets by the displaced papillary muscles, and of mitral annulus dimension, geometry, and function has been debated. ^21–25 The role of ventricular dysfunction with reduced transmitral pressure to close the leaflets has also been discussed and explored by echocardiography (transthoracic, transoesophageal, 3D, exercise) and videofluoroscopy. ^{9 , 11 , 27–29} Kwan et al ., using 3D-echocardiography, studied ischaemic and non-ischaemic cardiomyopathies and concluded that the mitral valve deformation was asymmetric in ischaemic disease and significantly more symmetrical in non-ischaemic cardiomyopathy. ³⁰ In that study, the mitral valve tenting area in the medial plane was the strongest determinant of severity of MR, in both ischaemic and non-ischaemic disease. The same relevance of mitral deformation indices on the degree of MR has recently been reported. ³¹ We observed that results too and we found a poor correlation between tenting height and LV diameters, volumes or its sphericity indices. Agricola et al . were stressing also the relevance of a precise measurement of valvular tenting area from parasternal long-axis view at mid-systole. Tenting area was the main determinant of MR severity. They studied the tenting and also the global intra-LV dyssynchrony with tissue velocity imaging. ²⁰ Dyssynchrony was also relevant for explaining MR severity. Our analysis provides complementary investigations, looking for not only the regional quantification of intra-LV dyssynchrony but also of related myocardial segment’s contractility. We were looking carefully to the mid-septum and also to the segment next to the anterolateral papillary muscle.

Role of dyssynchrony in the severity of MR

The mapping of LV function with strain imaging by Kanzaki et al . showed that the mechanical activation time delay between LV segments adjacent to the papillary muscles was associated with development of prominent MR. ¹⁴ They suggested that MR was related to a systolic imbalance of forces acting on the papillary muscles due to unco-ordinated regional LV activation in these segments, causing geometric changes in the mitral leaflet attachments. ¹⁴ This hypothesis is explaining the ability of CRT to decrease the degree of MR. ^{12 , 13 , 32} CRT appears to be able to co-ordinate the tethering forces on the papillary muscles and increase the leaflet coadaptational surface to reduce the amount of MR. This ‘acute’ mechanical effect was confirmed during temporary interruption of CRT in 25 patients, after 3 months of treatment. ³³ Similar observations were made in 21 patients treated with CRT for a median of 427 days. ³⁴ That last study demonstrated a CRT-dependant decrease in the degree of MR severity. As soon as the CRT device was turned off, the severity of MR increased, which was associated with a significant decrease in LV systolic function measured by d P /d t_max . ³³ This last observation might indicate that the severity of MR closely follows the degree of LV dyssynchrony as well as regional contractile dysfunction. The results of our study, using imaging of regional deformation, support this last hypothesis, but going one step further as we assessed the regional longitudinal contractility in the lateral wall and not only a global index of LV contractility. ³³ Thus, both dyssynchrony and intrinsic myocardial contractility have probably to be considered when discussing the therapeutic strategy of a CHF-patient with significant functional MR. Strain imaging appears highly relevant in such situation. It provides a unique tool to assess regional contractility and dyssynchrony, two relevant determinants of MR severity.

Limitations of our study

This observational mechanistic study was performed in consecutive patients with DCM and LV systolic dysfunction, regardless of the duration of their QRS. Our non-selected population provided an opportunity to describe the complexity of MR mechanisms. While the correlations, we observed, were somewhat weak, these results are, nevertheless, noteworthy, since they show that intervening on a single mechanism is unlikely to be curative in the management of MR. We were focusing on the LV lateral wall as this is usually the most delayed when an intra-LV dyssynchrony is observed. Nevertheless the recent contribution of Ypenburg et al . ³⁴ is also encouraging the use of strain data with regard to the posterior papillary muscle. Considering radial contraction instead of longitudinal one, they found that radial dyssynchrony with regard to the posterior papillary muscle was suggesting of an acute beneficial effect of cardiac resynchronization on MR severity.

Conclusions

Functional MR observed in patients suffering from CHF is multifactorial. LV regional systolic dysfunction and dyssynchrony appear to be both important determinants of MR severity. Their evaluation might require new echocardiographic or magnetic resonance imaging techniques to assess regional strain, since it might have therapeutic implications in the future. Our data suggest that analysing only LV geometry and mitral orifice is insufficient to correctly describe functional MR determinant.

Conflict of interest: none declared.

References