Prognostic implications of mitral valve geometry in patients with secondary mitral regurgitation: the COAPT trial 

Abstract

Aims

The impact of mitral valve geometry on outcomes after MitraClip treatment in secondary mitral regurgitation (MR) has not been examined. We therefore sought to evaluate the association between mitral valve geometry and outcomes of patients with heart failure (HF) and secondary MR treated with guideline-directed medical therapy (GDMT) and MitraClip.

Methods and results

Mitral valve geometry was assessed from the baseline echocardiograms in 614 patients from the COAPT trial. The primary endpoint for the present study was the composite of all-cause mortality or HF hospitalization (HFH) within 2 years. Effect of treatment arm (MitraClip plus maximally tolerated GDMT vs. GDMT alone) on outcomes according to baseline variables was assessed. Among 29 baseline mitral valve echocardiographic parameters, increasing anteroposterior mitral annular diameter was the only independent predictor of the composite endpoint of all-cause mortality or HFH [adjusted hazard ratio (aHR) per cm 1.49; P = 0.04]. The effective regurgitant orifice area (EROA) was independently associated with all-cause mortality alone (aHR per cm² 2.97; P = 0.04) but not with HFH, whereas increasing anteroposterior mitral annular diameter was independently associated with HFH alone (aHR per cm 1.85; P = 0.005) but not all-cause mortality. Other mitral valve morphologic parameters were unrelated to outcomes. MitraClip reduced HFH and mortality independent of anteroposterior mitral annular diameter and EROA (P_interaction = 0.77 and 0.27, respectively).

Conclusion

In patients with HF and severe secondary MR, a large anteroposterior mitral annular diameter and greater EROA were the strongest echocardiographic predictors of HFH and death in patients treated with GDMT alone and with the MitraClip.

Graphical Abstract

Open in new tabDownload slide

Introduction

In prior studies of patients with heart failure (HF) and secondary mitral regurgitation (MR), quantification of MR severity, indices of left ventricular (LV) function and remodelling, and clinical symptoms have been the principal parameters used to assess prognosis and select the most appropriate therapy.^1,2 Guideline-directed medical therapy (GDMT) for HF is foundational for all patients and may reduce MR by decreasing LV dimensions.^3,4 Cardiac resynchronization therapy has been shown to further reduce MR in some patients with HF.^5,6 When surgical coronary revascularization is needed, concomitant mitral annuloplasty is often performed.^1,2 However, recurrent MR after annuloplasty is not uncommon and has been associated with increased mortality and morbidity.^7,8 In patients with non-ischaemic cardiomyopathy and severe secondary MR, surgical mitral valve repair has also been associated with high operative risk and its variable durability.^7,8 Specific mitral valve geometric abnormalities and the extent of LV remodelling have been associated with recurrence of MR after surgical mitral valve repair and reduced survival.^7–10

More recently, transcatheter mitral valve repair (TMVr) has been introduced for the treatment of HF patients with severe secondary MR, the most widely studied of which is transcatheter mitral leaflet edge-to-edge approximation with the MitraClip. A decade ago, numerous mitral valve anatomic criteria were established that predicted success (or failure) of the MitraClip.^11,12 However, these studies were performed before widespread usage of this device. Procedural success rates have greatly improved within the last decade, and in the present era, less is known about the mitral valve anatomic parameters that predict procedural success and long-term outcomes with the MitraClip (and may vary in primary and secondary MR). In this regard, disparate outcomes of HF patients enrolled in the Cardiovascular Outcomes Assessment of the MitraClip Percutaneous Therapy for Heart Failure Patients With Functional Mitral Regurgitation (COAPT)¹³ and the Multicentre Study of Percutaneous Mitral Valve Repair MitraClip Device in Patients With Severe Secondary Mitral Regurgitation (MITRA-FR)¹⁴ trials were reported, which has most often been attributed to inter-study differences in LV dilation, LV function, and MR severity. However, a detailed analysis of mitral valve geometry and its interplay with LV remodelling and MR severity in determining the clinical outcomes of patients randomized to MItraClip plus GDMT vs. GDMT alone has not been performed in either study. Accordingly, in the present sub-study from the COAPT trial, we assessed the association between various geometric mitral valve parameters and clinical outcomes. We hypothesized that more advanced mitral valve apparatus deformation would be associated with worse outcomes independent of treatment arm, and that MitraClip treatment would mitigate the impact from these prognostic risk factors.

Methods

Study design and patient population

Details concerning the COAPT study design and patient population have been published previously.^13,15 In brief, HF patients with ischaemic or non-ischaemic cardiomyopathy and with moderate-to-severe (grade 3+) or severe (grade 4+) secondary MR who remained symptomatic despite the use of maximally tolerated GDMT were enrolled. Patients were randomized to TMVr using the MitraClip device (Abbott Vascular, Santa Clara, CA, USA) plus GDMT or GDMT alone. Procedural data have been published previously.¹³ The primary endpoint of the present analysis was the composite of all-cause death or HF hospitalization (HFH) at 24-month follow-up. Secondary endpoints included all-cause mortality and HFH alone. The protocol was registered on ClinicalTrials.gov (#NCT01626079) and was approved by the investigational board of each participating centre. All patients provided written informed consent. The sponsor participated in site selection and management and in data analysis. The principal investigators had unrestricted access to the data, wrote the manuscript, and vouch for the accuracy and completeness of the data and analyses and for the fidelity of the trial to the protocol.

Echocardiography

Baseline conventional echocardiographic measurements were performed by an independent echocardiographic core laboratory (MedStar Health Research Institute, Washington, DC, USA) and have been published previously.^13,16 Left and right ventricular dimensions and function were measured according to current recommendations.^17,18 MR was quantified using a multiparametric algorithm based on recommendations from the American Society of Echocardiography.^13,19 Tricuspid regurgitation (TR) was also graded based on a multiparametric approach consisting of qualitative and semiquantitative parameters.¹⁹

For the present post hoc analysis, mitral valve geometric analysis was performed using commercially available software for offline analysis (TomTec-Arena, version 20.18, 2017). Table 1 and Figures 1 and 2 summarize the mitral valve geometrical variables measured.^9,10,20–22 All measurements were done on transthoracic echocardiography. The measurements were performed per multiple beats if atrial fibrillation was present. The diameters of the mitral valve annulus were measured in the anteroposterior direction from the parasternal long-axis view and in the intercommissural direction from the apical two-chamber view. Mitral valve tethering was assessed measuring the distance between the coaptation point to septum, tenting height (the distance between the coaptation point of the mitral valve leaflets and the mitral valve annular plane), and tenting area (the area enclosed by the mitral valve leaflets and the mitral annular plane). The length of the anterior and posterior mitral valve leaflets were measured in mid-diastole on the parasternal long-axis view. The maximal tethering location was determined according to the scallops of the posterior mitral leaflet. The distance from the posterior papillary muscle to the mitral annulus was measured in the apical four- and two-chamber views, while the distance from the anterior papillary muscle to the mitral annulus was measured in the apical two-chamber view. The mitral posterior leaflet angle and the basal and distal anterior leaflet angles were measured on the apical long-axis view. Reproducibility of all the mitral valve echocardiographic measurements was tested in a randomly selected sample of 20 cases. Intraobserver agreement was very good or excellent for all mitral valve measurements [intraclass coefficient (ICC) >0.7], except for the length of the anterior and posterior mitral valve leaflets [ICC 0.5; 95% confidence interval (CI) 0.24–0.78 and ICC 0.57; 95% CI 0.15–0.74, respectively] and the anterior papillary muscle to mitral valve annulus distance (ICC 0.26; 95% CI −0.14 to 0.58). Inter-observer agreement was good, with ICC >0.6 for all measurements, except for the tenting height (ICC 0.39; 95% CI −0.05 to 0.63) and the coaptation to septal distance (ICC 0.45; 95% CI 0.09–0.71).

Statistical analysis

Categorical data are presented as frequencies and were compared by the χ² test or Fisher’s exact test, as appropriate. Continuous variables are presented as means and standard deviation or median (Q1–Q3) and were compared by Student’s t-test or Wilcoxon rank sum test. The independent association between the various clinical and echocardiographic parameters with the occurrence of the endpoint was assessed with univariate and multivariable Cox proportional hazards regression analysis. The relative effect of treatment vs. control arms on the risk of the endpoints associated with clinical and echocardiographic variables was assessed by formal interaction testing. Statistical analyses were performed with SAS version 9.4 (SAS Institute, Cary, NC, USA).

Results

Baseline clinical characteristics have been previously reported¹³ and are summarized in Supplementary data online, Table S1. The echocardiographic characteristics including the mitral valve geometry analysis are outlined in Table 2. The mean intercommissural (3.56 ± 0.43 cm) and anteroposterior (3.26 ± 0.39 cm) mitral annulus diameters were comparable demonstrating a circular rather than an elliptical shape of the mitral annulus. Significant tethering of the mitral valve leaflets was observed as indicated by the large distance between the point of coaptation and the interventricular septum, the increased tenting height/area, and the increased posterior leaflet angle. Leaflet tethering was most frequently observed at the central level of the posterior mitral leaflet (P2).

As previously described, patients had severe MR with a mean effective regurgitant orifice area (EROA) of 0.41 cm², regurgitant volume of 26.8 mL/beat, and regurgitant fraction of 36.4%. Moderately severe LV remodelling was observed with a mean LV end-systolic diameter of 5.3 cm and end-diastolic diameter of 6.2 cm. Mean LV end-systolic volume was 134.7 mL and end-diastolic volume 192.8 mL resulting in a mean LV ejection fraction of 31.3%. Mean right ventricular fractional area change was impaired with a mean value of 32.0%. The mean right ventricular systolic pressure (RVSP) was increased (44.3 ± 13.7 mmHg) and the majority of patients had mild TR.

Impact of mitral valve geometry and cardiac structure and function on MR reduction

Among 274 patients who were randomized to and treated with the MitraClip, at 30 days grade 0 or 1+ MR was present in 199 patients (72.6%), grade 2+ MR was present in 54 patients (19.7%), and grade 3+ or 4+ MR was present in 21 patients (7.7%). As shown in Table 3, patients with 3+ or 4+ MR at 30 days post-MitraClip had a significantly larger mitral posterior leaflet angle at baseline than those with ≤2+ MR. The tethering location was also significantly different between the groups; specifically, while mitral leaflet tethering at P3 was uncommon (7 patients), 3 of these patients (42.9%) had ≥3+ MR at 30 days after MitraClip. No other significant differences were observed in mitral valve geometry at baseline according to MR grade 30 days after MitraClip. In terms of conventional echocardiographic characteristics, patients with grade 3+ or 4+ MR 30 days after MitraClip had more severe MR at baseline, larger total stroke volume, larger LV end-diastolic volume index, and higher RVSPs.

Impact of mitral valve geometry on all-cause death or HFH at 24 months

In the entire patient population, larger baseline anteroposterior and intercommissural mitral annular diameters were the only mitral valve anatomic parameters that were significantly associated with the composite risk of death or HFH at 24 months (Table 4). Severe MR, moderate or greater TR, and increasing mitral EROA and RVSP were also significantly associated with the primary endpoint. By multivariable analysis, the baseline anteroposterior mitral annulus diameter was the only echocardiographic variable independently associated with the primary endpoint. Of note, baseline mitral EROA was not an independent predictor of the 2-year composite clinical outcome. A history of atrial fibrillation or flutter, New York Heart Association (NYHA) class and baseline B-type natriuretic peptide (BNP) plasma levels were also independently associated with occurrence of the primary composite endpoint within 2 years.

Death or HFH at 24 months occurred in 44.8% of patients randomized to MitraClip plus GDMT and in 66.1% of the patients randomized to GDMT alone (P < 0.0001). MitraClip treatment was an independent predictor of a 48% reduction in the primary composite endpoint (P < 0.0001; Table 4). By formal interaction testing, increasing anteroposterior mitral annular diameter was associated with a similar risk of the composite primary endpoint for both treatment arms (Figure 3A). Similarly, no significant interactions between atrial fibrillation, NYHA class or BNP plasma levels, and treatment arm were present for the primary composite outcome.

Impact of mitral valve geometry on all-cause death and heart failure hospitalization as separate outcomes

A larger mitral EROA was the only echocardiographic variable independently associated with an increased risk of all-cause mortality within 2 years (Supplementary data online, Table S2). Other independent correlates of increasing mortality were creatinine clearance, 6-min walk distance, BNP levels and EROA. MitraClip treatment was an independent predictor of a 38% reduction in 2-year mortality (P = 0.003). There were no significant interactions noted between these variables and treatment arm on 24-month all-cause mortality (Figure 3B).

Among the echocardiographic and mitral valve geometric variables, only the anteroposterior mitral annular diameter was independently associated with 2-year HFH (Supplementary data online, Table S3). Other independent correlates of increasing mortality were history of atrial fibrillation or flutter, NYHA class, and BNP levels. MitraClip treatment was an independent predictor of a 57% reduction in 2-year HFH (P < 0.001). There were no significant interactions noted between these variables and treatment arm on 24-month HFH (Figure 3C).

Discussion

In the present sub-study from the COAPT trial, among HF patients with severe secondary MR, an increasing anteroposterior mitral annular diameter was associated with the 2-year composite endpoint of all-cause mortality or heart failure hospitalization, an association that was driven by an increasing risk of HFH more so than mortality. In contrast, baseline EROA was independently associated with the 2-year rate of all-cause mortality but not HFH. Treatment with MitraClip compared with GDMT alone reduced the relative 2-year risks of all-cause death, HFH and the composite of death or HFH consistently in all subgroups, including larger vs. smaller anteroposterior mitral annular diameter and EROA. These findings demonstrate the complementary deleterious effects of a greater EROA (reflecting the severity of secondary MR) and an enlarged anteroposterior mitral annular diameter (reflecting long-standing volume overload). By reducing MR and having long-term favourable effects on LV remodelling,^13,16 treatment of severe secondary MR with the MitraClip plus GDMT thus improves freedom from both death and HFH during long-term follow-up in selected patients with HF.

Mitral valve geometry, MR severity, and clinical outcomes

LV remodelling in patients with ischaemic and non-ischaemic cardiomyopathy causes tethering of the mitral valve leaflets and reduced mitral valve closing forces.²³ Specifically, global and regional LV remodelling results in increased LV sphericity, papillary muscle displacement, mitral annular dilation, and planar flattening. Mitral valve closing forces are reduced due to impaired LV contractility, LV dyssynchrony, and reduced mitral annular contraction. The deleterious effects of mitral leaflet tethering and impaired mitral valve closing force leads to lack of leaflet coaptation. The deformation of the mitral valve apparatus may vary greatly depending on the extent of LV remodelling and dysfunction.²⁴ How the various alterations of the mitral valve geometry influence the outcomes of HF patients treated with MitraClip compared with GDMT alone has not been extensively investigated. Mantegazza et al.²⁵ showed that a severely enlarged anteroposterior mitral annular diameter (≥4.44 cm) was associated with significant residual MR after MitraClip implantation, but the authors did not report its effect on outcomes. The present substudy from the COAPT trial provides novel evidence by demonstrating an independent association between increasing anteroposterior mitral annular dilation and the composite 2-year outcome of all-cause mortality and HFH, and with the endpoint of HFH alone. Moreover, these associations were similar in both treatment arms suggesting that MitraClip plus GDMT will improve the relative prognosis across the range of anteroposterior mitral annular diameters enrolled within the COAPT trial, with the absolute benefits being greater with larger anteroposterior diameters.

Quantitative measures of MR severity (such as the EROA) have been associated with impaired outcomes. In secondary MR, an EROA >0.2 cm² has been associated with poor survival.¹⁹ However, randomized controlled trials have shown that surgical or TMVr techniques in patients with secondary MR have a prognostic benefit when the EROA (as a measure of MR severity) is at least 0.3 cm².^13,14,26

In the COAPT trial, in which the vast majority of patients had a baseline EROA ≥0.3 cm²,¹⁵ an increasing EROA was independently associated with increased risk of all-cause mortality alone but not with HFH. MitraClip reduced mortality to a similar relative degree regardless of baseline EROA; as such, the absolute reduction of mortality was greater in patients with greater severity baseline MR treated with the MitraClip.

Grayburn, Packer et al.⁸ have ventured that the MitraClip is especially likely to improve event-free survival in patients with secondary MR with an EROA that is disproportionate to the extent of LV remodelling. Many patients with proportionate MR are also likely to benefit from MR reduction. In contrast, patients with very advanced LV remodelling and marked dilatation may not respond to MR reduction with the MitraClip, even if the baseline EROA is abnormal. In COAPT, enrolment was strictly controlled to require 3+ or 4+ secondary MR (confirmed by an echocardiographic core laboratory prior to randomization), and LV dimensions were restricted by requiring an LV end-systolic diameter of <7 cm. This requirement likely limited the degree of mitral annular dilatation present in the patients enrolled. Whether the MitraClip would be effective if used in patients with greater LV remodelling and anteroposterior mitral annular diameters than studied with COAPT cannot be addressed by this study. Moreover, whether therapies such as mitral valve annuloplasty may be of greater benefit than the MitraClip or further improve outcomes if used in combination with the MitraClip when the mitral annulus is dilated deserves further study. Similarly, identifying when LV dilatation and mitral annular diameters are too abnormal to benefit from the MitraClip (or other therapies aimed at reducing secondary MR) is essential for optimal patient triage to TMVr or transcatheter mitral valve replacement vs. an LV assist device or heart transplantation.

Predictors of mitral regurgitation reduction after MitraClip

Procedural success rates with the MitraClip have steadily improved with increasing operator experience and advanced imaging guidance. In the present trial, 92.3% of MitraClip-treated patients achieved a reduction in baseline MR grade of 3+ or 4+ to <2+ at 30 days. Residual secondary MR ≥3+ has not surprisingly been shown in COAPT and other studies to be strongly associated with subsequent adverse events.^27–30 In a study of 300 patients with secondary MR undergoing MitraClip treatment, the EROA, mitral valve orifice area (MVOA), and transmitral pressure gradient were associated with inadequate MR reduction.²⁸ The early EVEREST II trial also excluded patients with LV end-systolic dimension >55 mm, MVOA <4.0 cm², and those with secondary MR and advanced tethering with either coaptation depth >11 mm or vertical coaptation length <2 mm.³¹ In contrast, COAPT excluded patients with degenerative mitral valve disease, LV end-systolic dimension >70 mm, MVOA <4.0 cm² and those who in the opinion of the site operator had leaflet anatomy that may preclude MitraClip implantation, proper positioning or sufficient reduction in MR; specific anatomic exclusion criteria were not otherwise provided.¹¹ Among patients meeting the enrolment criteria, we identified MR severity 4+ rather than 3+ (but not EROA), tethering location (especially P3), greater mitral posterior leaflet angle, stroke volume, LV end-diastolic volume index, and RVSP as predictors of greater residual MR after MitraClip implantation. Further analyses are underway to characterize the mechanisms and implications of these observations.

Limitations

As a post hoc analysis, the present results should be considered exploratory and hypothesis generating. The potential impact of unmeasured confounders cannot be excluded; thus, demonstrating associations in multivariable analysis does not prove causality. The operators participating in COAPT in general were very experienced; similar results may not be achieved by operators who have not yet traversed the learning curve.³² Not all echocardiographic parameters of interest were assessed in the present study, and additional insights might have been gained by 3D echocardiographic assessment.³³ Finally, by protocol patients with marked LV dysfunction were excluded, some pathologies (e.g. P3 tethering) were infrequently present, and of course, the results apply solely to treatment of functional MR. Further studies are thus warranted to examine the prognostic implications of mitral valve anatomies not included in the present study.

Conclusions

In the present echocardiographic core laboratory study from the multicentre COAPT trial, among HF patients with 3+ or 4+ secondary MR, a large anteroposterior mitral annular diameter was associated with increased risk of the composite outcome of all-cause death or HFH and HFH alone. Greater EROA was an independent predictor of mortality. Treatment with the MitraClip plus GDMT compared with GDMT alone reduced death and HFH consistently in patients with and without these extremes.

Supplementary data

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

References

Author notes