Author + information
- Received July 27, 2016
- Revision received November 1, 2016
- Accepted November 17, 2016
- Published online December 4, 2017.
- Alex Pui-Wai Lee, MDa,∗ (, )
- Chun-Na Jin, PhDa,
- Yiting Fan, MMa,b,
- Randolph H.L. Wong, MBChBc,
- Malcolm J. Underwood, MDb and
- Song Wan, MDb
- aDivision of Cardiology, Department of Medicine & Therapeutics, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong, China
- bCardiology Department, Renji Hospital, Medical College of Shanghai Jiao Tong University, Shanghai, China
- cDivision of Cardiothoracic Surgery, Department of Surgery, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong, China
- ↵∗Address for correspondence:
Dr. Alex Pui-Wai Lee, 9/F Lui Che Woo Clinical Sciences Building, Department of Medicine & Therapeutics, Prince of Wales Hospital, 30-32 Ngan Shing Street, Shatin, N.T., Hong Kong, China.
Objectives This study aimed to assess the hypothesis that mitral annular disjunction (MAD) is associated with abnormal annular dynamics due to decoupling of annular–ventricular function.
Background MAD, defined as a separation between the atrial wall–mitral valve (MV) junction and left ventricular (LV) attachment, is a structural abnormality occurring in MV prolapse (MVP). Few data exist on the 3-dimensional (3D) geometry of MAD and its functional implication.
Methods A total of 156 subjects including 101 MVP patients (58 ± 11 years), 30 subjects with normal MV (57 ± 15 years), and 25 heart failure patients with functional mitral regurgitation (66 ± 10 years) were studied using real-time 3D transesophageal echocardiography. The spatial relation between atrial wall, MV, and LV attachment was examined for MAD. The 3D extent of MAD and annular dynamics were quantitatively assessed. The LV global longitudinal strain and basal circumferential strains were measured by speckle tracking echocardiography.
Results MAD was evident in 42 MVP patients (42%), measuring 8.9 mm (6.3 to 10.7 mm), circumferentially spanning 87 ± 41°. Dynamically, normal and nondisjunctive annulus contracted and increased in a saddle shape during systole. In heart failure patients with functional mitral regurgitation, mitral annulus was dilated and relatively adynamic, probably related to poor LV function. In contrast, disjunctive annulus displayed paradoxical systolic expansion and flattening (p < 0.0001), despite preserved and comparable LV strains with normal patients. The 3D extent of MAD correlated significantly with abnormal annular dynamics and larger regurgitant orifice (p < 0.0001). In MVP patients without MAD, the LV global longitudinal strain correlated inversely with change in height (r = −0.61; p < 0.0001), whereas LV basal circumferential strain correlated with change in area (r = 0.61; p < 0.0001), but not in patients with MAD (p > 0.05).
Conclusions MAD is a common anatomic abnormality in MVP. The disjunctive annulus is decoupled functionally from the ventricle, leading to paradoxical annular dynamics with systolic expansion and flattening, and may thus require specific intervention.
Mitral valve (MV) prolapse (MVP) affects 3% of the population, and is the leading surgical indication for mitral regurgitation (1). Although leaflet degeneration has been considered the primary pathology, the importance of intrinsic annular abnormality in the pathogenesis of mitral regurgitation is increasingly recognized (2–5). The mitral annulus is a 3-dimensional (3D), saddle-shaped structure exhibiting dynamic conformational change in the cardiac cycle. The normal annulus contracts and increases in saddle-shaped nonplanarity during systole (6). Such annular dynamics are important for the balanced distribution of mechanical stresses imposed by the left ventricle (LV) on the MV. We have recently reported annular flattening in patients with MVP who developed greater mitral regurgitation than those with less regurgitation and normal individuals (4). Such annular flattening increases stress on the leaflets (4,7) and chordae (4), which can accelerate the degenerative processes. Intrinsic annular pathology in MVP is also suggested by dynamic studies (2,3,5). Nevertheless, the anatomic basis of abnormal annular structure and dynamics in MVP remains unresolved, and thus optimal treatment strategy is uncertain.
Normally, the base of the posterior MV leaflet joins the left atrial wall and the atrium–MV junction is connected to the LV on the atrial aspect of the ventricular free wall (8). The mitral annulus is in fact a tissue plane at the confluence of these structures. Motion of the annulus is, therefore, passively determined by contraction and relaxation of adjacent ventricular musculature and by motion of the aortic root (9). Mitral annular disjunction (MAD) is an anatomic abnormality of the annulus described pathologically as a wide separation between the atrium–MV junction and the LV attachment (Figure 1) that is appreciable on both gross and histologic examination (8). During MVP repair, surgical inspection may detect MAD as superior displacement or atrialization of the posterior leaflet base (10,11). Both transthoracic (12) and transesophageal (10,11) echocardiography can recognize MAD with excellent surgical correlation (Figure 1). However, few data exist on the 3D geometry of MAD and its relation with the annular and valvular function. Advances in real-time 3-dimensional echocardiography (RT3DE) allows quantitative analysis of the annular dynamics (2–5,13,14), whereas speckle tracking echocardiography can assess deformation of the LV musculature. We hypothesized that MAD is associated with abnormal annular function and mechanistically linked to decoupling of annular dynamics from ventricular deformation. To test this hypothesis, we used transesophageal RT3DE to study the 3D annular dynamics and valve morphology in relation to LV functional analysis assessed by speckle tracking echocardiography. We compared MVP patients with MAD with those without MAD and with normal subjects, as well as with patients with functional mitral regurgitation (FMR) due to heart failure. We sought to characterize the 3D anatomy of the disjunctive annulus and define its functional significance.
A total of 156 subjects, including 101 patients with MVP, 25 patients with FMR secondary to heart failure, and 30 control subjects referred for transesophageal echocardiography, were studied prospectively. MVP was identified on echocardiography as systolic displacement (>2 mm) of either or both mitral leaflets into the left atrium, beyond the mitral annular plane, as indicated in the long axis view. FMR was defined as MR secondary to LV systolic dysfunction (ejection fraction <40%) and/or annular dilatation without intrinsic MV abnormality. The indications for referral for transesophageal echocardiography included determination of regurgitation severity, evaluation of suitability of valve repair, suboptimal transthoracic images, exclusion of endocarditis, and evaluation of a cardiac source of an embolic event. Patients were excluded if they had contraindications to transesophageal echocardiography, mitral stenosis, aortic valve disease, pericardial or congenital diseases, endocarditis, or cardiomyopathy. All patients underwent 2-dimensional and 3D transesophageal echocardiographic examination using a standardized protocol for the evaluation of the leaflet and annular pathology. Patients with disjunction of the mitral annulus (MAD+ group), as defined below, were compared with those with no appreciable annular disjunction (MAD− group). Patients found to have no structural heart diseases on transesophageal echocardiography or arrhythmias were included as normal controls. The local institutional review board approved this study.
Transesophageal RT3DE of the MV was performed with an EPIQ7C ultrasound system (Philips Healthcare, Andover, Massachusetts) equipped with a fully sampled matrix transducer (X7-2t). Zoomed RT3DE image of the MV apparatus was acquired. The region of interest was adjusted to the smallest pyramidal volume that encompassed the entire MV to maximize frame rates (>15 Hz). Gated multibeat acquisition was performed carefully with patient breath-holding to avoid stitching artifact. The effective regurgitant orifice (ERO) of mitral regurgitation was quantified by the proximal isovelocity surface area method (15). For eccentric flow convergence, correction for flow constraint was performed by multiplying ERO by the ratio of the angle formed by the walls adjacent to the regurgitant orifice and 180°.
Detection and 3D characterization of annular disjunction
The volumetric datasets of the MV were analyzed offline on an Xcelera workstation (Philips Healthcare). The 3D annular anatomy was assessed quantitatively with dedicated software (QLAB 10.3, Philips Healthcare). Three orthogonal imaging planes were adjusted to ensure that the long axis planes bisected the MV and the short axis plane parallel to the plane of MV. The spatial relation of the posterior leaflet attachment to the left atrial wall and the LV basal myocardium was examined in multiple reconstructed radial planes at 10° intervals rotated around the long axis. MAD was present if there was a wide separation (≥5 mm) between the posterior leaflet insertion into the left atrial wall and the base of the LV free wall, based on original histologic description by Hutchins et al. (8) and echocardiographic description by Eriksson et al. (10). In patients with MAD (MAD+ group), the maximal disjunction distance among all imaging planes and the degree of disjunction arcs were measured. We summarized the disjunction extent by calculating the disjunction index, which was the product of the disjunction arc degree and the maximal disjunction distance (Figure 2). Mitral annular dimensions and shape were assessed using dedicated quantification software (Mitral Valve Navigator, Philips Healthcare) at both end-diastole and end-systole. Key annular parameters measured were annular intercommissural width, anteroposterior diameter, circumference, area, and height, as previously described (4). The ratio of annular height to intercommissural width was calculated for description of the degree of saddle shape of the mitral annulus normalized to annular size (7). A higher ratio indicates a deeper saddle shape. Dynamic change of each annular dimension was calculated by the formula:Leaflet dimensions and papillary muscle tip position were assessed quantitatively at end-systole, as previously described (4). Key parameters measured were leaflet area, billow height and volume, and lengths from papillary muscles to coaptation.
Left ventricular function
The LV ejection fraction, end-systolic and end-diastolic volumes, and left atrial end-systolic volumes were determined by volumetric analysis of the 3D datasets obtained by transthoracic RT3DE following the American Society of Echocardiography recommendations (16). Two-dimensional speckle tracking echocardiography was used to evaluate LV strains (17). The left ventricular global longitudinal strain (LV-GLS) was analyzed offline (QLAB 10.3, Philips Healthcare) by averaging the peak strains derived from apical views, and the left ventricular basal circumferential strain (LV-BCS) from the basal short axis view.
Data are expressed as mean ± SD, median (interquartile range), or number of patients (percentages) as appropriate. Normality of continuous data was analyzed with the Shapiro–Wilk test. All continuous variables except ERO, leaflet billow volume, and disjunction distance and index were normally distributed. Group comparisons for baseline characteristics, summarized annular measurement (systolic/diastolic average), valve measurements at specific time points, and LV strains used analysis of variance (ANOVA), Kruskal-Wallis test, chi-square test, or Fisher exact test as appropriate. Pairwise comparisons after a significant ANOVA or Kruskal-Wallis test were made by post-hoc Tukey honest significant difference or Steel-Dwass tests, respectively. Intragroup comparisons of each annular measure between end-systole and end-diastole used a paired Student t test. Repeated-measures ANOVA analyzed differences in the dynamic changes of annular measurements from end-diastole to end-systole between groups. Correlations between normally distributed variables were tested by Pearson analysis. Correlations of disjunction index with other variables were tested by Spearman coefficient. Intraobserver and interobserver variabilities of 3D measurements of MAD geometry were assessed using intraclass correlation coefficient of repeated analysis in 10 randomly selected subjects by the same operator 1 week apart, and by a different operator. Reproducibility of other 3D MV echocardiographic measurements was reported previously (4). Analyses were performed with JMP version 12.0 (SAS Institute Inc., Cary, North Carolina). A value of p < 0.05 was considered significant.
Table 1 shows patient characteristics. Of the 101 subjects with MVP, MAD was seen in 42 patients (42%). A minor degree of annular disjunction (<5 mm) was seen in 2 normal control subjects (one 28-year-old man and one 59-year-old woman). MAD was not seen in the FMR group. The number of prolapsed segments was significantly greater in MAD+ than MAD– group (3.8 ± 4.3 vs. 1.2 ± 2.2; p < 0.0001). Segmental distribution of lesions was different (chi-square = 54.8; p < 0.0001), with isolated P2 and/or P3 prolapse predominant in MAD– group, and bileaflet prolapse in the MAD+ group. Conversely, chordal rupture was more frequent in the MAD– group (chi-square = 4.5; p = 0.034). The LV and left atrial volumes were increased significantly in both MVP groups when compared with normal subjects. There were no differences between the MAD+ and MAD– groups regarding age, sex, LV volumes, ejection fraction, left atrial volume, and ERO. The FMR patients were significantly older (p < 0.05 vs. the other 3 groups). The LV volumes of the FMR groups were significantly more dilated with lower ejection fraction, but the left atrial volume and the ERO were smaller when compared with the MVP groups (all p < 0.05).
Location and 3D geometry of MAD
The intraclass correlation coefficients for intraobserver and interobserver variability of disjunction distance measurement were 0.96 and 0.82, respectively; that of disjunction arcs measurement were 0.98 and 0.81. In general, MAD is located adjacent to the prolapsed segments, circumferentially spanning over 87 ± 41° (range 20° to 170°) of the annulus, with the maximal disjunction distance measuring 8.9 mm (6.3 to 10.7 mm). The maximal disjunction was located most frequently at P2 (n = 20) and P1 (n = 18), less commonly in P3 (n = 4). In the 2 normal subjects showing minor MAD, the separation distances were 3.4 and 4.9 mm (both <10°) at P2 and P1, respectively.
Relation of annular disjunction with annular structure and dynamic function
Averaged measures of annular dimensions are shown in Table 2. Compared with normal subjects, the mitral annulus of both MVP groups had increased anteroposterior diameter, intercommissural width, area, and circumference (all p < 0.0001 for both MVP groups vs. normal). Intercommissural width (p = 0.0083) and area (p = 0.039) were significantly greater in MAD+ than MAD− group, but anteroposterior diameter (p = 0.32) and circumference (p = 0.054) were similar. Annular height was similar in both MAD+ and MAD− groups (p = 0.20), but the annular height-to-intercommissural width ratio were significantly lower (p = 0.0058) in the MAD+ group. The FMR annulus was dilated significantly, with increased intercommissural width, anteroposterior diameter, circumference, and area (all p < 0.05 vs. normal). The intercommissural width and circumference were enlarged to an extent greater than that of both MVP groups (all p < 0.05). The annular height of the FMR group was similar to controls, but higher than the MAD+ group (p = 0.012).
Dynamically, dramatic intergroup differences were observed (Table 2, Figure 3). In normal subjects, the annular anteroposterior diameter, intercommissural width, circumference, and area decreased from end-diastole to end-systole (all p < 0.0001). The annular saddle deepened in systole with increased height and height-to-intercommissural width ratio (both p < 0.0001). In MAD− patients, the annular anteroposterior diameter, circumference, and area decreased in systole (all p < 0.05), but was less than normal (p < 0.05). The intercommissural width (p = 0.91) and height (p = 0.53) remained unchanged, meaning that the annulus did not exhibit significant deepening of its saddle shape in systole. In contrast, in MAD+ patients, the anteroposterior diameter, intercommissural width, and area paradoxically increased in systole (all p < 0.005). In addition, the height and height-to-intercommissural width ratio paradoxically decreased (both p < 0.0001), resulting in paradoxical annular “unsaddling” in systole. In FMR patients, the intercommissural width, circumference, and area decreased from end-diastole to end-systole (all p < 0.0001); the anteroposterior diameter (p = 0.29) and annular height (p = 0.19), however, remained unchanged. Systolic increase of the annular height-to-intercommissural width ratio (p = 0.0003) was, therefore, contributed mainly by contraction of the intercommissural width rather than an increase in the annular height in FMR patients. In MAD+ patients, the disjunction index correlated significantly with ERO (ρ = 0.56; p = 0.0001), change in intercommissural width (ρ = 0.63; p < 0.0001), change in area (ρ = 0.60; p < 0.0001), change in circumference (ρ = 0.49; p = 0.001), change in height (ρ = −0.50; p = 0.0007), and change in height-to-intercommissural width ratio (ρ = −0.59; p < 0.0001). There is a trend of correlation between disjunction index and change in anteroposterior diameter (ρ = 0.28; p = 0.07).
Relation of annular disjunction with valvular structure and function
As shown in Table 3, MAD was associated with more severe mitral leaflets and chordae tendineae deformity. Compared with normal and MAD− groups, MAD+ group had significantly larger leaflet areas, billow height and volume, and longer lengths from papillary muscle to coaptation (all p < 0.05).
Relation of annular disjunction with left ventricular function
Despite normal LV ejection fraction, the LV-GLS of both MAD+ (20.0 ± 4.7% [p = 0.014] vs. normal) and MAD− (−19.2 ± 3.4% [p = 0.0003] vs. normal) groups was significantly (ANOVA p = 0.0005) reduced when compared with normal controls (−22.9 ± 2.2%). The LV-GLS was comparable between MAD+ and MAD– patients (p = 0.54). The LV-BCS was similar across groups (MAD+ vs. MAD– vs. normal = −21.2 ± 5.6% vs. −20.9 ± 5.3% vs. −20.3 ± 4.6%; ANOVA p = 0.81). In MAD– patients, the LV-GLS correlated inversely with change in height (r = −0.61; p < 0.0001), whereas the LV-BCS correlated with change in area (r = 0.61; p < 0.0001). In contrast, LV strains did not correlate with any annular dynamics in MAD+ patients (p > 0.05).
To the best of our knowledge, the current study is the first to characterize the 3D structure of MAD and its functional significance. We found that MAD is relatively common in MVP. The disjunctive annulus displayed paradoxical systolic dilatation and flattening and is associated with more diffuse leaflet deformity. Such abnormal annular dynamics occur despite relatively normal LV systolic function assessed by both ejection fraction and strains, suggesting intrinsic annular abnormalities and a decoupled annular function.
Normal mitral annular structure and dynamics
The normal mitral annulus undergoes complex conformational changes during the cardiac cycle. Such motions are not inherent to the fibrous annulus, but are consequences of external forces exerted by adjacent ventricular musculature (18). Torrent-Guasp proposed in his helical band model that systolic shortening of the circumferential basal loop fibers produces strain to reduce annulus dimensions like a sphincter (18,19). Systolic deepening of annular saddle is also a passive motion produced by differential translational annular movements (2,4,7). Longitudinal LV fiber contraction translates the posterior annulus apically. The anterior annulus, tethered at the aortic root (9), has to tilt posteriorly, thus folding the annulus into a saddle shape. In addition, LV longitudinal contraction pulls the medial and lateral annulus downward and inward, adding curvature. When the annulus is disjunctive, its motion no longer follows LV contraction, but exhibits paradoxical dynamics, conforming to atrial wall motion. The annulus simply stretches out with left atrium during systole (Figure 4). Our study provided quantitative evidence of annuloventricular coupling by showing a close relation between the annular conformational dynamics with ventricular strains in normal and nondisjunctive annulus. The absence of such a relation in MVP patients with disjunctive annulus can be explained by functional decoupling of the annular and ventricular motions that entails anatomic disjunction. These findings support that MAD has important functional implications.
Patients with FMR have a significantly dilated annulus. Although the degree of annular enlargement in FMR is comparable with that of MVP patients with MAD, the annular structure and dynamics significantly differ between the 2 groups. MAD was not seen in FMR patients. In contrast with the paradoxical annular expansion seen in MAD, the FMR annulus contracts in systole. The fractional changes of circumference and area are lesser than normal; anteroposterior diameter and height are essentially adynamic. These findings are consistent with previous studies (2,3) and are probably related to poor ventricular function.
Our observations support the role of annuloplasty in MVP repair; annular abnormality is common. Recent studies highlighted the potential advantage of saddle-shaped annuloplasty to restore annular shape (20,21). However, our findings imply that implantation of an annuloplasty ring from the atrial side may not restore the function of disjunctive annulus fully because annular–ventricular decoupling may persist (22). Refinement of annuloplasty techniques by taking into consideration the 3D geometry of disjunction and ventricular fiber orientation may be important for sustained surgical outcome (10,11). Similarly, transcatheter MV repair techniques that target leaflet pathology (e.g., MitraClip) may not correct fully the annular dysfunction associated with disjunction, and thus the long-term outcome of such therapy in this subset of patients warrant further investigation.
Our study provides new insights into the pathophysiology of MVP. Autopsy findings of MAD led Hutchins et al. (8) to propose that regional disjunction permits hypermobility of atrium–valve junction. In the present study, we used 3D echocardiography to confirm the functional importance of MAD by demonstrating its association with paradoxical annular dynamics. Systolic annular dilation may accentuate mitral regurgitation. Indeed, the extent of disjunction in the MAD+ patients correlates with the degree of mitral regurgitation. Furthermore, paradoxical systolic annular unsaddling may exaggerate mechanical stresses on the leaflets and chordae, which may predispose and perpetuate leaflet/chordal degeneration. In the MAD− phenotype, despite annular dilatation, annuloventricular coupling remains intact. Intriguingly, longitudinal LV strain is reduced in MVP and severe mitral regurgitation, possibly due to LV remodeling. This may suggest that, in these patients, the primary pathology lies in leaflet and chordal tissue deficiency, whereas annular dysfunction, although present, may only be secondary.
From our observation, a lesser degree (<5 mm) of separation between the atrial–valve junction and its ventricular attachment can be seen occasionally in patients with otherwise structurally normal MV. Two control subjects were noted to have such minor separation. The separation exists in only 1 imaging plane (i.e., circumferentially spanning <10° of the annulus). A minor degree of annular disjunction has been described in autopsy of normal hearts (8). Such minor “disjunction” may represent a normal anatomic variant or an early or milder form of pathological disjunction. Indeed, it would be intriguing to understand whether these minor disjunctions will progress and predispose leaflet prolapse. However, there are not enough of these patients in our study population to draw any conclusion about such an observation. The clinical significance of minor disjunction warrants further longitudinal follow-up echocardiographic studies.
The findings of the present study support the prevalence of intrinsic annular abnormality in patients with MVP. For the first time in humans, our study shed lights on the structural basis for this functional abnormality by demonstrating the decoupling of annular and ventricular function in relation to anatomic annular disjunction. The comprehensive description of MAD geometry in this study has laid down the framework for future works on new surgical techniques and devices. Our results also challenge clinicians to think of the MV as an integrated mechanism susceptible to valvuloventricular interactions.
COMPETENCY IN MEDICAL KNOWLEDGE: MAD occurs in 42% of patients with MVP undergoing transesophageal echocardiography. It represents an intrinsic annular abnormality that leads to paradoxical annular enlargement and flattening during systole. The extent of annular disjunction is related positively to the degree of mitral regurgitation. Such dynamic abnormalities of the annulus may be related to decoupling of the annular and ventricular functions.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: Evaluation of the presence and extent of MAD should become an essential part of pre-procedural planning of surgical and transcatheter MV repair.
TRANSLATIONAL OUTLOOK: Additional studies are needed to determine the prognostic importance of MAD in terms of the progression of primary mitral regurgitation. Further research can be conducted to explore whether a minor degree of annular disjunction in otherwise normal MVs may progress and predispose future valve prolapse. The therapeutic value of treatment strategies targeting to correct intrinsic annular structural abnormality such as disjunction can be evaluated in further studies.
This work was partially supported by the General Research Fund of the Research Grant Committee (467812), Hong Kong, China. Dr. Lee has received research equipment support and speaker honorarium from Philips Healthcare. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- effective regurgitant orifice
- functional mitral regurgitation
- left atrium
- left ventricle
- left ventricular global longitudinal strain
- mitral annular disjunction
- mitral valve
- mitral valve prolapse
- real-time 3-dimensional echocardiography
- Received July 27, 2016.
- Revision received November 1, 2016.
- Accepted November 17, 2016.
- 2017 American College of Cardiology Foundation
- Grewal J.,
- Suri R.,
- Mankad S.,
- et al.
- Levack M.M.,
- Jassar A.S.,
- Shang E.K.,
- et al.
- Lee A.P.,
- Hsiung M.C.,
- Salgo I.S.,
- et al.
- Clavel M.A.,
- Mantovani F.,
- Malouf J.,
- et al.
- Salgo I.S.,
- Gorman J.H. 3rd.,
- Gorman R.C.,
- et al.
- Lee A.P.,
- Acker M.,
- Kubo S.H.,
- et al.
- Mor-Avi V.,
- Lang R.M.,
- Badano L.P.,
- et al.
- Ormiston J.A.,
- Shah P.M.,
- Tei C.,
- Wong M.
- Buckberg G.,
- Hoffman J.I.,
- Mahajan A.,
- Saleh S.,
- Coghlan C.