Differences in mitral valve geometry between atrial and ventricular functional mitral regurgitation in patients with atrial fibrillation: a 3D transoesophageal echocardiography study

Abstract

Aims

This study investigated geometric differences in mitral valve apparatus between atrial functional mitral regurgitation (A-FMR) and functional mitral regurgitation (FMR) with left ventricular (LV) dysfunction in patients with atrial fibrillation (AF) using 3D transoesophageal echocardiography (TOE).

Methods and results 

In total, 135 moderate or greater FMR patients with persistent AF or atrial flutter underwent 3D TOE. Fifty-six patients had A-FMR, defined as preserved LV ejection fraction (LVEF) of ≥50% and normal LV wall motion. Seventy-nine patients had ventricular FMR (V-FMR), defined as LV dysfunction (LVEF of <50%) or LV wall motion abnormality. To evaluate mitral leaflet coaptation, the coapted area was calculated as follows: total leaflet area (TLA) in end-diastole − closed leaflet area in mid-systole. Although annular area (AA) did not significantly differ between the two groups, TLA was significantly smaller in A-FMR than in V-FMR (P = 0.005). TLA/AA, indicating the degree of the leaflet remodelling, was significantly smaller in A-FMR than in V-FMR (P < 0.001). A-FMR had significantly smaller posterior mitral leaflet tethering height and angle measured at three anteroposterior planes (lateral, central, and medial) than V-FMR (all P < 0.001). However, vena contracta width (VCW) measured on long-axis view on TOE and coapted area, which correlated with VCW (r = −0.464, P < 0.001), were similar between the two groups.

Conclusion

Mitral leaflet remodelling may be less in A-FMR compared with V-FMR. However, leaflet tethering was smaller in A-FMR than in V-FMR, and this may result in a similar degree of mitral leaflet coaptation and mitral regurgitation severity.

Introduction

Functional mitral regurgitation (FMR) is commonly attributed to local or global left ventricular (LV) dysfunction and remodelling without structural mitral valve (MV) abnormalities. Local or global LV dysfunction and remodelling cause papillary muscle (PM) displacement and mitral leaflet tethering resulting in reduced leaflet coaptation.^1,2 Long-standing atrial fibrillation (AF) may result in left atrial (LA) enlargement and mitral annular dilatation. Mitral annular dilatation plays an important role in the development of atrial FMR (A-FMR).^3–5 Recently, other factors, such as insufficient leaflet remodelling, posterior mitral leaflet (PML) tethering, decreased annular contractility, and flattening of the annular saddle shape, were reported to be associated with A-FMR development.^6–10 However, the precise mechanism of A-FMR remains controversial.

With the unique capability of real-time 3D transoesophageal echocardiography (TOE), accurate information of MV apparatus can be obtained. Over the years, the MV geometry between patients with A-FMR and non-significant A-FMR and/or healthy individuals with sinus rhythm has been compared using 3D TOE.^6,7,9,10 To the best of our knowledge, only few echocardiographic studies have focused on the difference in MV geometry between A-FMR and FMR with LV dysfunction in patients with AF. To identify potential therapeutic targets and to develop appropriate treatment strategies using an approach different from that used in FMR with LV dysfunction, the precise mechanism of A-FMR must be elucidated. Therefore, this study investigated the geometrical differences in MV apparatus between A-FMR and FMR with LV dysfunction in patients with AF using 3D TOE.

Methods

Participants

In total, 260 consecutive moderate or greater FMR patients with persistent AF or atrial flutter (AFL) underwent 3D TOE from September 2014 to August 2019 at our institution. FMR was defined as mitral regurgitation (MR) without structural MV abnormalities (including degenerative change, stenosis, rheumatic disease, congenital mitral anomaly, vegetation, mass, and history of surgical or transcatheter intervention). Persistent AF or AFL was defined as one detected on electrocardiogram at the time of TOE and lasting longer than 7 days before TOE. The exclusion criteria were as follows: (i) congenital heart disease (n = 1), (ii) history of surgery for aortic valve (AV) (n = 19), (iii) history of transcatheter AV replacement procedure (n = 11), (iv) moderate or greater AV disease (n = 48), (v) use of LV assist device (n = 4), (vi) severe mitral annular calcification (n = 2), (viii) missing transthoracic echocardiography (TTE) data within 3 months before or after TOE (n = 28), and (viii) inadequate TOE image quality for 3D analysis (n = 12). Among the patients who underwent TTE after TOE, those who only had TTE data after cardioversion for AF (n = 2), catheter ablation for AF (n = 1), percutaneous MV intervention (n = 6), and surgical MV intervention (n = 6) were excluded. These TTE data were included in missing TTE data, one of the exclusion criteria. The remaining AF or AFL patients (n = 135) were classified into two groups: 56 patients with A-FMR, defined as preserved LV ejection fraction (LVEF) of ≥50% and normal LV wall motion and 79 patients with ventricular FMR (V-FMR), defined as LV dysfunction (LVEF of <50%) or LV wall motion abnormality (Figure 1). The study protocol was approved by the review board of Cedars-Sinai Institution (no. 00000814).

Two-dimensional echocardiography

A comprehensive 2D and Doppler TTE was performed using an ultrasound system (iE33; Philips Andover, MA, USA) by experienced sonographers. TOE was performed using an ultrasound system (iE33 or EPIQ7; Philips) with the X7-2t or X8-2t ultrasound probe while the patients were sedated. Thus, blood pressure was maintained during the examination. All echocardiographic images were stored on a PICOM365 (ScImage, Inc., Los Altos, CA, USA) and reanalysed offline by a single investigator blinded to the clinical data. All measurements were evaluated based on the current guidelines.^11–14 LV volume and LVEF were assessed using the modified Simpson’s method from apical two- and four-chamber views.¹² LA volume was measured from standard apical two- and four-chamber views using biplane area-length method at end-systole.¹² LV and LA volumes were indexed to the body surface area.

We assessed the underlying mechanism and severity of valvular heart diseases using both TTE and TOE. The MR severity was evaluated using the semiquantitative and quantitative parameters including vena contracta width (VCW), effective regurgitant orifice area, and regurgitant volume based on the proximal isovelocity surface area method.^11,14 We measured VCW as the narrowest width of the jet measured from the long-axis view of the TTE image and mid-oesophageal TOE image. The MR severity was graded as mild (1+), moderate (2+), moderate to severe (3+), and severe (4+).

Three-dimensional echocardiography and analyses

The 3D datasets of the entire MV were obtained using the live 3D zoom mode or one-beat full-volume mode under optimal depth and gain setting. All 3D volumetric datasets were analysed offline using a commercial software [mitral valve navigation (MVN) and cardiac 3D quantification (3DQ), QLAB, version 13.0, Philips]. We selected one representative frame in mid-systole and another in end-diastole to assess MV geometry. The mid-systole is the frame when the mitral leaflet coaptation shifts most closely to the LA and the end-diastole is the last frame prior to mitral leaflet coaptation.

The MVN is originally designed for systolic MV assessment. However, we used this tool to determine in diastole the total leaflet area (TLA) as in the previous study.⁶ We adapted the MVN to measure the TLA accurately in diastole when both leaflets were open. When the mitral leaflet area was measured in diastole using the MVN, we identified the tips of the anterior mitral leaflet (AML) and PML and traced both leaflets separately. However, the mitral leaflets could not be traced accurately when the mitral leaflets were fully opened and the angle between the mitral annulus and leaflet was >90°. Therefore, we selected the end-diastole frame when the leaflets were not fully opened.

Using the MVN, we measured the 3D parameters of MV in both mid-systole and end-diastole as follows (Figure 2): (i) three orthogonal planes were displayed and three images, including bi-commissural, long-axis, and short-axis views, were manually adjusted. (ii) Four annular reference points, including anterior, posterior, anterolateral, and posteromedial points, were marked on long-axis and bi-commissural views. (iii) The mitral annulus shape was manually outlined by editing multiple annular points on eight image planes rotated around the axis perpendicular to the mitral annular plane. (iv) In mid-systole, the coaptation points between the AML and PML were marked, and the closed mitral leaflets were then manually traced in multiple parallel long-axis planes from commissure to commissure. In end-diastole, the tips of the leaflets were marked, and the whole AML and PML were respectively traced in multiple parallel long-axis planes from commissure to commissure. (v) The MVN automatically reconstructed a 3D model of the MV apparatus, and geometric measurements were generated, which included MV annular anteroposterior (AP) diameter, anterolateral–posteromedial (ALPM) diameter, height, circumference, annular area (AA), closed leaflet area, defined as the closed leaflet surface to the LA, tenting volume, and tenting height in mid-systole and TLA, which was the summation of the AML and PML areas in end-diastole. The tenting volume was indexed to the body surface area. The annular height-to-ALPM diameter (annular height/ALPM diameter) ratio, an indicator of the annular saddle shape, was calculated. The ratio of TLA in end-diastole divided by AA in mid-systole (TLA/AA) was calculated as the degree of leaflet remodelling relative to the mitral annulus.^6,8,15,16 To assess the degree of coaptation between AML and PML, the coapted area was calculated as TLA in end-diastole minus closed leaflet area in mid-systole.

The 3D measurements of PML tethering height and angle were determined on three AP planes (lateral, central, and medial) perpendicular to the commissure–commissure plane in mid-systole using 3DQ and MVN (Figure 3).¹⁷

Statistical analyses

Continuous variables were expressed as mean ± standard deviation. To assess the differences between A-FMR and V-FMR, statistical analyses were performed using unpaired Student’s t-test or Mann–Whitney U test for continuous variables, as appropriate. Categorical data were presented as numbers (percentages) and compared between groups using the χ² test. Paired Student’s t-test or Wilcoxon-signed rank tests were used to compare haemodynamic parameters and body weight between TTE and TOE. The correlation between continuous variables was investigated via a simple regression analysis and expressed as Spearman’s correlation coefficient. The reproducibility of 3D TOE measurements was evaluated in 14 randomly selected patients (n = 7, A-FMR and n = 7, V-FMR) using intraclass correlation coefficient. For intraobserver variability, the same observer performed the measurements again 3 months after the initial assessment. For interobserver variability, a second observer selected the same 3D datasets and performed 3D analyses independently. Statistical analyses were performed using the Statistical Package for the Social Sciences software (version 23.0, SPSS Inc., Chicago, IL, USA). P-values < 0.05 were considered statistically significant.

Results

Clinical characteristics of participants

Clinical characteristics are shown in Table 1. There was a significantly higher proportion of female patients in A-FMR than in V-FMR (P = 0.003). V-FMR had a significantly lower systolic blood pressure and a higher prevalence of class III or IV symptoms based on the New York Heart Association classification and coronary artery disease than A-FMR (all P < 0.05). In terms of laboratory data, the B-type natriuretic peptide level was significantly higher in V-FMR than in A-FMR (P = 0.001).

Two-dimensional echocardiographic findings

Haemodynamic parameters, such as systolic blood pressure, diastolic blood pressure, and heart rate, and body weight, were similar between TTE and TOE (Supplementary data online, Table S1). Table 2 shows the 2D echocardiographic findings. LVEFs were 60% ± 6% and 32% ± 11% in A-FMR and V-FMR, respectively. V-FMR had a significantly higher LV end-diastolic volume index and LV end-systolic volume index than A-FMR (all P < 0.001). However, no significant difference was observed in the LA volume index between the two groups. Additionally, systolic pulmonary artery pressure and the prevalence of tricuspid regurgitation (moderate or greater) did not differ significantly between the two groups. However, V-FMR had a significantly lower tricuspid annular plane systolic excursion than A-FMR (P = 0.017).

MR severity

The MR severity is shown in Table 2. There was no significant difference in the MR grade between A-FMR and V-FMR. In addition, no significant differences were observed in effective regurgitant orifice area, regurgitant volume, and VCW measured on the long-axis view of TTE and TOE images between the two groups.

Three-dimensional TOE analyses

MV annular parameters

The MV annular parameters between A-FMR and V-FMR are shown in Table 3. No significant differences were observed in MV annular parameters.

MV area parameters and leaflet remodelling

The MV area parameters obtained via 3D TOE are shown in Table 3 and Figure 4a. Although there was no significant difference in AA between the two groups, the closed leaflet area and TLA were significantly smaller in A-FMR than in V-FMR (P = 0.003 and 0.005, respectively; Figure 4a). The scatter plot showed a correlation between TLA and AA. TLA had a strong positive correlation with AA in A-FMR and V-FMR (r = 0.957 and 0.942, respectively, both P < 0.001) (Figure 4b). The comparison of TLA/AA between A-FMR and V-FMR is shown in Table 3 and Figure 4c. TLA/AA was significantly smaller in A-FMR than in V-FMR (P < 0.001).

MV leaflet tethering

A-FMR had a smaller tenting volume (P < 0.001) and tenting height (P < 0.001) than V-FMR (Table 3). The PML tethering height and angle in three parallel AP planes were significantly smaller in A-FMR than in V-FMR (all P < 0.001; Table 3and Figure 5).

Coapted area and MR severity

The coapted area was comparable between the two groups (Figure 6a), which indicated that the degree of leaflet coaptation was similar. To evaluate the correlation between the coapted area and MR severity assessed using VCW measured on the long-axis view of the mid-oesophageal TOE image in 135 FMR patients with AF or AFL, simple regression analyses were performed (Figure 6b). A negative correlation was observed between the coapted area and VCW in all patients with FMR (r = −0.464, P < 0.001).

Representative cases

The representative cases are presented in Figure 7. Panels (a) and (b) depict the representative cases of A-FMR and V-FMR, respectively. TLA/AA in panel (a) was lower than that in panel (b) (1.18 vs. 1.29), indicating panel (a) had less leaflet remodelling than panel (b). Additionally, panel (a) had a smaller tenting height than panel (b) (3.4 vs. 9.5 mm), showing that panel (a) had relatively flat leaflets than panel (b). However, the coapted area was similar between panels (a) and (b) (112 vs. 115 mm²), indicating that the degree of leaflet coaptation was similar between panels (a) and (b).

Reproducibility

Intraobserver and interobserver variabilities assessed using intraclass correlation coefficient (95% confidence interval) for 3D measurements were as follows: AA: 0.97 (0.91–0.99) and 0.95 (0.50–0.99), TLA: 0.96 (0.88–0.99) and 0.93 (0.56–0.98), coapted area: 0.82 (0.53–0.94) and 0.79 (0.46–0.93), tenting volume: 0.93 (0.81–0.98) and 0.83 (0.57–0.94), and tenting height: 0.90 (0.72–0.97) and 0.87 (0.65–0.96), respectively.

Discussion

As far as we know, this is the first study to investigate the geometrical differences in MV apparatus between A-FMR and V-FMR in patients with AF using 3D TOE. The major findings were as follows: (i) TLA/AA, indicating the degree of leaflet remodelling, was smaller in A-FMR than in V-FMR. (ii) A-FMR had less mitral leaflet tethering than V-FMR. (iii) The degree of leaflet coaptation and MR severity were similar between A-FMR and V-FMR.

Geometrical differences in leaflet remodelling between A-FMR and V-FMR

Long-standing AF causes LA enlargement resulting in mitral annular dilatation. A-FMR is commonly characterized by dilated mitral annulus.³ However, there is no consensus as to whether the condition alone causes A-FMR.^3,4,18 In this study, no significant differences were observed in both LA size and mitral annular area between A-FMR and V-FMR. Therefore, mitral annular dilatation is a common geometrical change in A-FMR and V-FMR. In contrast, TLA was significantly larger in V-FMR than in A-FMR. Previous experimental and clinical studies have shown mitral leaflet expansion in response to mitral annular dilatation and leaflet tethering in ischemic and dilated cardiomyopathy.^15,16,19,20 This apparent compensatory mechanism may occur as leaflet remodelling in order to prevent MR. A strong positive correlation was observed between TLA and AA in both patients with A-FMR and V-FMR, indicating that leaflet remodelling as a compensatory mechanism occurs in both A-FMR and V-FMR. This finding was in accordance with that of previous studies.^6,8,20 Insufficient leaflet remodelling has been associated with the development of significant FMR with LV dysfunction in a previous 3D TOE study.²⁰ Regarding A-FMR, recent studies have revealed that insufficient remodelling may be associated with the development of A-FMR.^6,8 Thus far, the comparison of leaflet remodelling between A-FMR and V-FMR has not been reported. In this study, we found that TLA/AA was significantly smaller in A-FMR than in V-FMR. This result may be explained by a previous study, which showed that the mechanical stresses caused by mitral leaflet tethering can cause leaflet expansion.¹⁶ In other words, the leaflet relative expansion or remodelling in A-FMR was not as large as in V-FMR because of less tethering seen in A-FMR.

Geometrical differences in leaflet tethering between A-FMR and V-FMR

In FMR patients with LV dysfunction, the degree of mitral leaflet tethering had a positive correlation with MR severity.^2,21 PM displacement caused by LV dysfunction and remodelling led to restricted leaflet motion and distant coaptation point of the leaflet from the mitral annular plane. As for A-FMR, the feature of the leaflet coaptation and the degree of tethering have remained controversial. Previous studies have shown that patients with A-FMR had a similar or smaller tenting height than controls.^5,6 However, other studies showed that A-FMR had a tethered PML.^7–9 These controversial findings can be attributed to the fact that there are different methods for evaluating leaflet tethering, including height and angle, and also anatomical points to measure, including leaflet coaptation or tip. Therefore, we chose to measure both tethering height and angle in both A-FMR and V-FMR.

Degree of leaflet coaptation and MR severity

The degree of mitral leaflet coaptation is important for the assessment of FMR. However, the degree of mitral leaflet coaptation cannot be easily assessed because each leaflet involved in coaptation cannot be visualized accurately. Previous studies have investigated the coaptation index for the assessment of mitral leaflet coaptation in patients with FMR using the following formula: coaptation index = (TLA at the onset of MV closure − closed leaflet area in mid-systole)/TLA at the onset of MV closure × 100.^5,22 Although the leaflet area at the onset of mitral leaflet closure in systole was calculated as TLA, it can be difficult to specify the onset of MV closure and to measure TLA accurately in systole when there is an overlapping area of AML and PML.^6,8,15,16,19 Therefore, we measured TLA in end-diastole when both leaflets could be visualized separately. As far as we know, this is the first study to determine the coapted area as the subtraction of total leaflet surface area to LA measured in systole (closed leaflet area) from the summation of both leaflets area measured in diastole (TLA) with 3D TOE. Our study showed that the coapted area had a negative correlation with MR severity. The coapted area and MR severity in A-FMR were similar to those in V-FMR. This similar coapted area may be explained by the less expansion or remodelling of the leaflets with less tethering in A-FMR.

Clinical implications

Considering the geometrical differences between A-FMR and V-FMR, surgical mitral annuloplasty may be more suitable or successful in patients with A-FMR, which has less leaflet tethering, than in those with V-FMR.

Limitations

The current study had several limitations. First, this was a single-centre retrospective research. Additionally, we only included patients who underwent 3D TOE for the assessment of MV. Therefore, patients in our study may not represent the generalized population in the community. Second, the sample size was relatively small. However, to the best of our knowledge, the number of participants in our study was the largest among others that performed 3D TOE. Third, the temporal sequence of LV function based on our classification of A-FMR and V-FMR was not considered. Thus, V-FMR might have potentially included patients who originally should have been classified under A-FMR and developed later LV dysfunction due to volume overload caused by chronic A-FMR. Thus, future studies should focus on the temporal change in MV geometry among patients with A-FMR who developed LV dysfunction caused by chronic MR. Finally, identifying the true leaflet tips in diastole was occasionally difficult in patients with calcified mitral leaflets and/or chordae tendineae. In those patients, leaflet tips were carefully identified by checking other frames before and after the selected one. In this study, fortunately, we measured only FMR patients who did not have heavy calcification of the leaflets and/or chordae tendineae. Therefore, we believe that identifying the leaflet tips is relatively accurate.

Conclusions

Geometrical differences exist in leaflet remodelling and tethering between A-FMR and V-FMR in patients with AF. Mitral leaflet remodelling may be less in A-FMR compared with V-FMR. However, A-FMR has less leaflet tethering than V-FMR, and this may result in a similar degree of mitral leaflet coaptation and MR severity. Considering the anatomical difference, A-FMR may require different therapeutic strategies from V-FMR.

Acknowledgements

The authors thank Shinsuke Ito, Toshiyuki Onodera, and Takayuki Ogura from Philips Japan, for their technical assistance.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Data availability

The data pertaining to this article will be shared on reasonable request to the corresponding author.

Conflict of interest: none declared.

References