Echocardiographic assessment of ischemic mitral regurgitation

Accurate grading of MR is central to clinical decision-making. MR should be graded using an integrative approach, incorporating multiple Doppler techniques for direct quantification as well as supportive data (left atrial size, LV chamber size, pattern of pulmonary vein flow) in the overall assessment [11]. Color Doppler techniques include:

where R is the radius of the hemispheric PISA zone (Figure 6)

The PISA method provides a quantitative method for MR grading. However, the calculation requires a geometric assumption of a hemispherical shape to the PISA region which is not always the case. Additionally, it can be technically challenging to measure the PISA radius accurately.

Tables 3 and 4 shows reference ranges for color Doppler criteria for MR grade based on 2003 American Society of Echocardiography guidelines [11]; however 2014 American College of Cardiology/American Heart Association guidelines propose a new classification scheme of valvular disease severity, based on a combination of echocardiographic and symptomatic parameters, with stages of “at risk” to “progressive” to “asymptomatic severe” to “symptomatic severe” [2]. Recent consensus statements also endorse lower cut-off values for EROA for CIMR severity as compared to primary MR. In part this is due to 1) data that shows worse prognosis at smaller EROA in CIMR, likely reflecting the effects of the incremental volume load of lesser degrees of MR on an already dysfunctional ventricle, and 2) 2D echocardiographic underestimation of the flow convergence-method derived EROA due to “crescentic” orifice geometry in CIMR as opposed to a circular orifice [2].

In addition to semi-quantitative and quantitative Doppler techniques, it is important to integrate supportive and complementary data into the overall severity grading. Pulmonary venous flow reversal is specific for severe MR although of lower sensitivity (Figure 7). Chamber enlargement (LA and LV), dense continuous wave MR Doppler profile, and elevated E wave peak velocity >1.2 m/s are all suggestive of severe MR [11‐13] (Figures 8 and 9).

3D echocardiography has been demonstrated to provide accurate and reproducible MR grading using 3D guided planimetry of the VC area, which is essentially equivalent to the direct measure of the EROA. An advantages of 3D measurement of the EROA is that it does not require geometric assumptions that are used for 2D EROA calculation. Disadvantages are the lower frame rates of 3D color Doppler, which can effect lateral resolution and hence may erroneously exaggerate the area measured [14].

Finally, CIMR is a dynamic process, and the echocardiographer must consider how ambient preloading and afterloading conditions such as patient’s volume status, systemic blood pressure, and medications may affect the observed degree of MR.

Quantitative measures have attempted to correlate LV systolic dysfunction and LV dilatation with CIMR. Elegant experimental observations show that isolated LV systolic dysfunction (pharmacologically induced in a large animal model) does not produce significant MR [6, 15]. This is likely because without tethering forces, relatively little closing force is required to be generated by the LV to force the mitral leaflets toward the annular coaptation zone. However, in the same model of pharmacologic LV systolic dysfunction, when the LV was allowed to dilate by relieving an extrinsic pericardial restraint, MR was generated. This observation confirms as a key mechanism the apical and outward dilatation of PMs which create tethering forces on the leaflets.

In CIMR with symmetric tethering, LV end-systolic and end-diastolic volumes and the sphericity index correlate with the severity of MR. This is because the degree of LV dilatation directly relates to apical displacement of the PMs. For asymmetric tethering phenotypes the measures of global LV remodeling do not as robustly correlate with severity of MR because a small infarct can disrupt PM geometry and generate severe MR; the actual measures of mitral valve deformation are better predictors (see below). LV dilatation would therefore not be an independent predictor of CIMR severity in a population with mixed CIMR phenotypes.

The normal orientation of the PMs is with their long axis parallel to that of the LV and perpendicular to the plane of the mitral annulus. A local infarct that disrupts myocardium underlying a PM can radically change the relationship of that PM relative to the other PM and to the valve apparatus. This asymmetric effect of the infarct on the posteromedial PM translates directly into creating asymmetry in mitral valve apparatus anatomy and function – by rotating the posteromedial PM, tethering the posterior leaflet, and deforming the posterior portion of the mitral annulus – which creates a substrate for eccentric CIMR (Figure 3 and Additional file 1). Several lines of experimental and echocardiographic evidence correlate post-infarct inferoposterior wall motion abnormality with severity of MR [16]. Direct evidence that PM displacement generates CIMR was obtained in a sheep study of echocardiography-guided PM repositioning by an inflatable balloon external to the myocardium [17]. In this study, a Dacron patch with an adjustable balloon was sewn epicardially over areas of infarct after circumflex artery ligation; inflation of the balloon could be tailored to reduce the ischemic dilatation of the inferior wall, thus re-approximating PM geometry, and reducing severity of MR without a change in measures of LV contractility.

In practice, echocardiographic measurement of PM displacement requires intracardiac landmarks. The anterior mitral annulus is anchored at the aortomitral fibrous curtain, and this point in the parasternal long axis or apical four chamber views can provide a reference for measurement of apical displacement of both PM heads (Figure 10B,C) [18]. In a population of 128 LV systolic dysfunction patients, the strongest multivariate correlations with MR severity in a functional MR model were the apical displacement of the posteromedial PM and the inferoposterior displacement of the anterolateral PM [18]. In the parasternal short axis view at the mid-ventricular level, PM body displacements may be referenced relative to the mathematical center of the LV. Agricola and colleagues constructed a “mid-septal perpendicular line,” bounded by the septal insertions of the right ventricle myocardium, from which to measure posterior displacements of the PMs (Figure 10D) [7]. Lateral displacements of both PMs were measured from a second line constructed orthogonal to the mid-septal perpendicular line. Finally, a distance between the papillary body muscles was recorded. Regardless of CIMR phenotype, absolute value of each of these displacement measures is higher when compared to normal controls. In addition, the displacement measures will tend to be higher in symmetric versus asymmetric CIMR, but the magnitude of the changes between phenotypes is a few millimeters and thus not sufficient to differentiate them without other information on mitral valve deformation (Table 5). Some differences correlate with asymmetric phenotypes, e.g. the ratio of posterior displacements of the posteromedial:anterolateral PMs is about 1.2 in asymmetric CIMR but about 0.94 in symmetric CIMR or normal controls [7]. 3D TTE permits additional insight into the geometric angles relating both PMs to the LV cavity long axis, with greater asymmetry in the angles in CIMR versus functional MR with a dilated cardiomyopathy [19]. 3D TTE can be used to measure true spatial vector distances from the aortomitral curtain to the PM tips [20] and also characterize the spatial geometry of the PMs in relation to the annulus [21].

Wall motion abnormalities are critically important in gauging local LV dysfunction in CIMR: the echocardiographer should identify and quantify wall motion as part of a comprehensive assessment of a global assessment of ischemic burden. Indices of wall motion abnormalities underlying the posteromedial PM insertion are highly important in assessing CIMR. Novel methodologies, including LV basal rotation dynamics as assessed by speckle tracking, further highlight local differences between myocardial function in symmetric and asymmetric phenotypes [10]. Normal systolic rotation may contribute to decreasing the distance from posterior PM head to leaflet and mitral annular contraction. In a multivariable model, impairment of basal rotation was a key predictor of CIMR severity after inferoposterior MI, likely because of less ability of myocardial rotation to reduce adverse tethering lengths and also a contribution to reduction of mitral annular contraction.

Ischemic and/or systolic PM dysfunction itself does not seem to contribute to CIMR on top of the contribution of PM displacement. Kaul first reported poor overall correlation of reduced PM thickening and MR severity in canines [24]. In a sheep model of CIMR by left circumflex occlusion but with preserved PM blood supply via a perfusion catheter from the aorta, withdrawal of the perfusion catheter caused onset of papillary ischemia as measured by decreased strain rate but was correlated with diminished tethering distances and reduced MR [25]. In humans, there is some evidence that PM dysfunction, as measured by longitudinal systolic strain, actually reduces MR observed after inferior myocardial infarction [26]. Impairment of PM contraction presumably reduces tension on the chordae and paradoxically compensates for the tethering forces exerted by PM misalignment and/or LV dilatation. Novel protocols employing delayed enhancement cardiac magnetic resonance imaging confirmed that while PM infarct was observed in 30% of patients at 4 weeks after first myocardial infarction, neither partial nor complete PM infarct robustly correlated with CIMR [27]. These observations reinforce the notion that geometric PM displacement, and not necessarily systolic function, is the key factor in determining CIMR.

The aggregate of the abnormal vector forces on the mitral leaflets manifest echocardiographically as incomplete mitral leaflet closure or tenting; as such it represents the common pathway of LV remodelling and PM displacement in CIMR. Various measures of quantifying tethering and tenting are available by routine 2D TTE techniques. The incomplete mitral leaflet closure pattern is often best appreciated in the apical four chamber view, because the mitral annular plane is defined in this view.A single linear measure of “tenting height” – the maximal mid-systolic distance from mitral leaflet tips to the annular plane – reflects the abnormal apical shift of the coaptation zone (Figure 10A). While this measure has been correlated with CIMR severity, tenting height understandably may be different when the tethering forces are directed posterolaterally versus apically for example because height alone it does not account for angle of tethering relative to the annular plane.

The tethering angles define the relationship of the base of the leaflets to the annulus: α represents the angle between annular plane and anterior mitral leaflet and β the angle between annular plane and posterior mitral leaflet [23]. 3D TTE and TEE acquisition of volumetric data sets allows selection of particular imaging slices to calculate tethering angles [22]. Though the exact values depend on methodology and imaging plane selected, higher ratios of posterior angle to anterior angle characterize the asymmetric tenting phenotypes, and also predicts increased MR severity [22].

Tenting area provides a more integrative measurement that is less dependent on a particular angle, and also accounts for the geometry of the entire leaflet and not just that at the annular attachment. Tenting area is calculated as the area bounded by the anterior and posterior leaflets and the mitral annular plane (Figure 10); this measurement is performed at mid-systole, when the tenting area would be at a maximum. In the VALIANT-Echo substudy of 341 patients with echocardiographic LV ejection fraction <35% after myocardial infarct, tenting area was the only independent predictor of progressively worsening CIMR based on followup TTE data to a median 24.7 months [28]. Tenting area above a threshold of 4 cm² predicted near 6 fold odds of having moderate or greater MR at the end of followup and an odds ratio of 3.6 for increase in the degree of MR. In patients with LV systolic dysfunction, tenting area was a major determinant of functional MR severity, independent of global LV function, LV volume, and spherical shape. Tenting area itself correlates with linear measures of apical or posterior PM displacements [18]. Extending the analogy of assessment of tenting beyond tenting height and tenting area, tenting volume as defined by 3D echocardiography affords another level of comprehensive measurement of mitral valve deformation. However, the importance of the tenting phenotype must be considered, because even with the same indices of tenting height, area, or volume, an asymmetric CIMR phenotype will likely be associated with more significant MR (Figure 11).

Finally, secondary chordal attachments (basal or strut chordae) to the anterior mitral valve leaflet may exert additional geometric constraints on systolic MV configuration, most commonly manifesting as a bend, or an angle, between the distal and basal portions of the anterior mitral leaflet that further impairs coaptation. This angle can give a qualitative visual clue, assessed as a convexity or concavity in the configuration of the anterior mitral valve leaflet toward the left atrium in the parasternal long axis view in systole, with concavity indicating a bowing into the LV that strongly correlated with CIMR severity [29].

The mitral annulus has a specialized 3D geometry likened to an ovoid saddle shape that reduces stresses on the leaflets and supports valvular competence [30]. Dilatation of the annulus may occur secondary to either LV or LA dilatation, and while dilatation occurs primarily along the posterior annulus, even the fibrous anterior portion of the mitral annulus may dilate [31, 32]. Additionally, dilatation along the posterior annulus may be asymmetric, with a predilection for the region of the posterior commissure (P₂ – P₃ segment).

Annular dilatation can cause an incomplete coaptation pattern due to insufficient available leaflet area. However, the degree of dilatation does not necessarily correlate with the severity of CIMR. Distortion of the native 3D annular geometry to a “flattened” annulus may also contribute to CIMR by changing the leaflet motion. However in a study of lone atrial fibrillation patients with annular dilatation but normal LV chamber size, significant MR was not observed [33]. This is because LV remodelling and dilatation is required to generate tethering forces, though the study did show a weak correlation between functional MR severity and annular area.

Annular dilatation can be measured by anterior and posterior dimensions, annulus area (apical four chamber mitral annulus dimension multiplied by apical two chamber mitral annulus dimension multiplied by π/4) and perhaps with more computationally sophisticated methods such as the MVQ software package (Mitral Valve Quantification, Phillips). Surveillance of mitral annular dilatation is a part of our practice because of a self-propagating cycle of annular dilatation → MR → LV dilatation → annular dilatation. Mitral annular contraction, equal to (diastolic annular area – systolic annular area)/diastolic annular area, has a negative correlation with MR severity in LV systolic dysfunction [18] and in post-infarct MR.

Work by Robert Levine at Massachusetts General Hospital has described 3D echocardiographic methods to compare the areas of the mitral leaflets to the “closing area” and the annular area [34, 35]. In human models of functional MR, the mitral leaflet areas are greater than in patients without dilatation or prior infarct. However, the ratio of the measured mitral leaflet area to the calculated “closing area” is decreased in functional MR. There may a threshold lower ratio which would be consistent with diagnosing a functional MR mechanism; it may be possible in the future to echocardiographically detect, measure and monitor this process as a means of assessing the remodelling response to CIMR. The biologic response that allows the valve to remodel by enlargement and thickening seems to be due to reactivation of embryonic development pathways occurring within the leaflet tissue [36].

The mechanisms responsible for recurrence of CIMR after surgical revascularization and restrictive annuloplasty remain elusive. In some instances, the mechanism is ongoing adverse LV dilatation and spherical remodelling that worsens tethering [37, 38]. In a single center retrospective population of predominantly ischemic MR, preoperative LV end diastolic diameter indexed to body surface area with a cut-off of >3.5 cm/m² predicted recurrence of MR [39]. A greater degree of anterior mitral leaflet tethering angle α, specifically >36.9° (considered the moderate-to-severe or severe quintiles of anterior tethering), regardless of LV dilatation or geometry, conferred a multivariate OR of 3.6 for recurrent MR at 44.7 month follow-up of CIMR patients who underwent surgical revascularization and undersized ring annuloplasty [40]. This is in accord previous results showing α ≥39.5° conferred OR of 3.1 for recurrent MR in a similar population of patients who underwent surgical revascularization and undersized ring annuloplasty [41]. There was also a strong association (OR >4) for lack of LV reverse remodeling post-operatively. The results of this line of analysis underscores that preoperative echocardiography and tethering geometry does predict postoperative outcomes including MR recurrence, LV geometry, and outcome, and thus these should be part of preoperative assessment. Preoperative diastology may also impact postoperative outcome, with transmitral deceleration time <140 ms predictive of MR recurrence, and deceleration time and pulmonary vein systolic:diastolic flow ratio predictive of mortality [42].

Because annuloplasty shifts the coaptation zone more anteriorly, the posteromedial PM location can be further distorted and lie outside the annulus ring; the tethering effect on the posterior leaflet makes it less likely to coapt at the anteriorly shifted coaptation zone [43]. In patients without continued global LV dilatation, recurrent MR is highlighted by adverse anterior leaflet tethering due to bending, as measured by anterior leaflet coaptation area [38].

Derangements in peak systolic longitudinal, radial, and circumferential strain measures mirror underlying wall motion abnormalities in both asymmetric or symmetric CIMR [13]. In symmetric CIMR, peak systolic strain was reduced globally, while in asymmetric CIMR phenotypes there was more localized systolic strain derangements in the inferoseptal and inferior territories. While it is not yet clear how strain might add to the diagnosis of CIMR, it could assume a particular role in surgical planning: in a 61 patient CIMR cohort, strain did not improve after surgical revascularization and restrictive annuloplasty in the symmetric group, but did improve at one year in the asymmetric group [13].

Patients with mild (or “progressive”) rest CIMR may exhibit more severe inducible regurgitation as assessed by flow convergence methods [13], and this may represent the etiologies of exertional symptoms [44] and excess mortality seen with CIMR [45]. Exercise physiology exerts multiple effects that bear on the mitral valve apparatus and degree of MR: inotropy increases which augments global and regional LV systolic dysfunction and has the potential to improve mitral valve coaptation geometry; against this, exercise contributes to increased LV systolic pressure and increased chronotropy with shortened systolic time, which contribute to augmented transmitral LV to left atrial pressure gradient [46]. Additionally, exercise-induced ischemia could contribute to new or worsened WMA and tethering, or increased heart rate and altered loading conditions may result in worsening of ventricular mechanics which in setting of underlying akinesis or dyskinesis, result in increased MR. The net change in ischemic MR with exercise depends ultimately on which factor(s) represent the underlying mechanism of the ischemic MR: about one-quarter of CIMR patients show decrease ischemic MR with exercise [47], e.g. those with inferior myocardial infarction who can augment LV function with exercise and who would not have worsened ventricular mechanics.

Exercise may present a method to risk stratify patients with LV systolic dysfunction and mild rest CIMR at rest, as cardiovascular mortality at 19 month followup was predicted by worsening of mild rest CIMR (judged by an increase in EROA ≥13 mm² on a symptom-limited semisupine bicycle exercise test for which beta blockers were held for 24 hours) [47]. In another study of submaximal Bruce protocol treadmill exercise with patients on beta blockers, no rest echocardiographic parameters predicted severity of exercise-induced CIMR by EROA; instead only changes in exercise-induced mitral geometry measured by valve tenting area and coaptation distance represented the independent predictors of ischemic MR severity [48]. Exercise echocardiography may be reasonable in patients with ischemic heart disease and suspected CIMR who report dyspnea disproportionate to rest MR and/or LV dysfunction or who experience pulmonary edema without explained cause, and for whom additional information would answer whether surgery would benefit [13, 46].

TEE can be a useful adjunct to TTE for characterizing the mechanism of MR (particularly for intrinsic leaflet pathologies) and mapping anatomic defects. It may help exclude an organic etiology when assessing the patient with CIMR, and also provide better spatial resolution of chordal and leaflet geometric relationships. The use of TEE intraoperatively and post-operatively in the evaluation of MR has been comprehensively reviewed by Sidebotham et al. [49] and Shakil et al. [50]. TEE is important in assessment of patients undergoing surgical revascularization as it provides another opportunity to assess for CIMR. However, because of vasodilating effects of anesthesia, CIMR severity may be underestimated by intraoperative TEE. One proposed tactic to ensure appropriate severity grading is to administer vasopressors to mimic more physiologic afterload conditions. In a single study, the proxy for physiologic afterload was a systolic blood pressure of 160 mmHg though the exact target is debatable; concurrent with vasopressor administration, most patients’ pulmonary artery occlusion pressure rose and only a few patients were administered extra intravenous fluid to combat the venodilating effects of anesthesia [51].

Non-echocardiographic cardiac imaging modalities are being deployed to study CIMR. These techniques may require the patient to remain immobile and flat and to perform breath holds – potential issues for patients with orthopnea due to cardiomyopathy or MR. Computed tomography implies a radiation exposure and magnetic resonance may require significant time as well as specialized equipment. Nevertheless, robust data sets with axial and three dimensional information may be derived which are suitable for a comprehensive classification of the interwoven geometry of the components of the mitral valve apparatus, for example augmented definition of annulus dimensions, annulus height, shape, and tenting height and angles [31, 52, 53] Delayed enhancement cardiac magnetic resonance and CT also offer alternative routes to more precise definition of region of PM and LV myocardial infarct [27], and are thus useful to establish the underlying ischemic etiology of MR and also define myocardial viability which may impact treatment decision-making [2]. The role of computed tomography and magnetic resonance remains to be defined.