Abstract
Background
In secondary MR, data on left ventricular (LV) remodeling after MitraClip procedure are rare, even this information may impact patient selection. This study investigated changes in LV structure and function by cardiovascular magnetic resonance (CMR) following MitraClip implantation for secondary mitral regurgitation (MR) in order to assess extent and predictors of LV reverse remodeling (LVRR).
Methods and Results
Twenty-nine patients underwent CMR imaging prior to and six months after MitraClip procedure. LVRR was defined by a decrease of LV end-diastolic volume index (LVEDVi) > 15% compared to baseline. According to the definition of LVRR, 34% of patients displayed LVRR at follow-up CMR. Baseline LV stroke volume index (LVSVi), LV ejection fraction (LVEF), LV circumferential strain and MR volume at baseline were predictors of LVRR at follow-up. At second CMR, we detected an improvement in hemodynamic status as illustrated by an increase in effective LVSVi (28 ± 8 ml/m2 vs. 33 ± 8 ml/m2; p = 0.053) and cardiac index (2.0 ± 0.5 vs. 2.3 ± 0.5 l/min; p = 0.016), while LVEF and LV strain parameters did not change (p > 0.05). Improvements in effective LVSVi were associated with the decrease of MR volume (r = 0.509; p = 0.018) and MR fraction (r = 0.629; p = 0.002) by MitraClip.
Conclusions
Together, MitraClip implantation is associated with LVRR in one third of patients. Baseline LV function and magnitude of MR are important predictors of LVRR. Improvement of hemodynamic status may be assessed by effective stroke volume index and correlates with the reduction of MR by MitraClip implantation, rather than an increase in LV contractility.
Keywords: Mitral regurgitation, Transcatheter mitral valve repair, Cardiovascular magnetic resonance, Left ventricular remodeling
Abbreviations
- AUC
Area Under the Curve
- CAD
Coronary Artery Disease
- CMR
Cardiovascular Magnetic Resonance
- Effective LVSVi
Effective Left Ventricular Stroke Volume Index
- HF
Heart Failure
- LGE
Late Gadolinium Enhancement
- LV
Left Ventricle
- LVEDVi
Left Ventricular End-diastolic Volume index
- LVEF
Left Ventricular Ejection Fraction
- LVRR
Left Ventricular Reverse Remodeling
- LVSVi
Left Ventricular Stroke Volume Index
- MR
Mitral Regurgitation
- ROC
Receiver Operating Curve
- RV
Right Ventricle
- TMVR
Transcatheter Mitral Valve Repair
- TR
Tricuspid Regurgitation
1. Introduction
Transcatheter mitral valve repair (TMVR) with the MitraClip device has evolved to an established treatment option for patients with clinically relevant mitral regurgitation (MR) and elevated surgical risk. While immediate procedural success rates reached 90% and the majority of patients clinically benefits from TMVR, long-term mortality remains substantial and appears to be related to prognosis of underlying heart failure (HF) [1], [2], [3].
The effect of TMVR on left ventricular (LV) remodeling is highly variable [4]. Previous echocardiographic studies reported LV reverse remodeling (LVRR) in a fraction of about 50% of patients undergoing TMVR with the MitraClip device in various patient cohorts [5], [6], [7], [8]. However, especially in patients with secondary MR, clinical benefit as well as ventricular response following TMVR are less clear. In addition, data on predictors of LVRR after TMVR and its impact on clinical outcome are still limited, though this information may impact patient selection for MitraClip. So far, only few, small studies with mixed cohorts implemented serial cardiovascular magnetic resonance (CMR) scans in patients referred for MitraClip [9], [10].
Here, we evaluated the impact of MitraClip implantation on ventricular volumes and function assessed by sequential CMR imaging in order to determine extent and predictors of LVRR in secondary MR.
2. Materials and Methods
Forty patients that underwent MitraClip implantation at the university hospital Duesseldorf, Germany were screened between 2014 and 2018. In eight patients CMR imaging was not possible (claustrophobia n = 3; refused CMR scan n = 3; obesity n = 2). Thus, 32 patients were scheduled for CMR imaging prior to and six months after the procedure. Three patients died prior second CMR scan and were excluded, so that the final patient́s cohort consisted out of 29 patients that all underwent serial CMR scans prior and after TMVR. Patients that were included had severe secondary MR and were considered at high surgical risk by an interdisciplinary heart team. Logistic EuroScore was used for risk stratification. All patients underwent comprehensive cardiological assessment including coronary angiography, right heart catheterization, transthoracic and transesophageal echocardiography prior to the procedure. The study was approved by the ethics committee of the Heinrich-Heine University Duesseldorf (study number 6110R) and executed in accordance with the Declaration of Helsinki. Patients were stratified into two groups according to the presence of LVRR. LVRR was defined as previously described by a decrease of LV end-diastolic volume index (LVEDVi) > 15% at second CMR compared to baseline [6], [11].
CMR was conducted with a 1.5 Tesla MRI scanner (Achieva, Philips Medical Systems, Best, The Netherlands) using a 32-channel phased array coil. Functional and structural assessment was determined by cine steady state free precession (SSFP) images in standard long axis geometries (two-, three- and four-chamber view) as well as in short axis orientation. Flow and velocities in the ascending aorta and pulmonary trunk were determined by using through plane velocity encoded imaging in a retrospective gating technique during one single breath hold. LV and right ventricular (RV) volumes were assessed by manually tracing endocardial borders at end-diastole and end-systole in each of the short-axis cine images. In case of atrial fibrillation, three beats were averaged. LV and RV end-diastolic volume and end-systolic volume were assessed using the slice summation method and matched to body surface area to calculate LV and RV end-systolic and end-diastolic volume indices (LVESVi/RVESVi/LVEDVi/RVEDVi). The LV and RV stroke volume index (LVSVi/RVSVi) was the difference between LVEDVi and LVESVi as well as RVEDVi and RVESVi. Effective LVSV reflecting the LV forward flow into the aorta was calculated by the difference of total aortic forward flow minus aortic backward flow and indexed to body surface area (effective LVSVi). In 27 patients, gadolinium-based contrast agent (ProHance®, Bracco Imaging) was given for assessment of late gadolinium enhancement (LGE). Two patients did not receive contrast agent due to reduced renal function. Post processing analyses were performed offline using commercial software (cmr42, Circle Cardiovascular Imaging Inc., Calgary, Alberta, Canada and Extended Workspace, Philips Healthcare, Hamburg, Germany). Strain analysis was accomplished using dedicated software (Image-Arena Version 3.0 and 2D Cardiac Performance Analysis MR Version 1.1.0; TomTec Imaging Systems Unterschleissheim, Germany).
Patients were scheduled for a second CMR scan six months after the MitraClip procedure. All-cause mortality, HF-hospitalizations and New York Heart Association (NYHA) class III/IV were assessed during 12 months follow-up. The clinical course was monitored by follow-up examinations, phone calls to the referring cardiologists and the patients‘ primary physicians or the patients themselves.
2.1. Statistical analysis
All analyses were performed using SigmaPlot (Systat Sotware Inc., San Jose, California, USA) and Graphpad Prism (Graphpad Software, San Diego, USA). Data for continuous variables are presented as mean ± SD or median with interquartile range. Continuous variables were tested by the Kolmogorov-Smirnov test to assess normality of distribution. Categorical variables are presented as frequencies and proportions. Differences between two groups were compared for significance with a two-tailed unpaired t-test. Fisheŕs exact test was used to examine the significance of the association between two kinds of classification. Simple logistic regression analysis was performed to assess predictors of LVRR at follow-up. Multivariate analysis was not performed due to low number of events. Receiver operating curve (ROC) analysis was constructed to evaluate strongest predictors of LVRR and Youdeńs Index for optimal thresholds. Correlation between the change of MR volume and MR fraction and the change in effective LVSVi was assessed by Pearson correlation. For all analyses, a p-value of < 0.05 was considered to be statistically significant.
3. Results
The final patient cohort included 29 patients that underwent CMR imaging prior and 5 ± 2 months following MitraClip implantation. Baseline patient characteristics are shown in Table 1. The medication of the patient cohort during the study period is given in Table S1 and Figure S1. Mean age was 77 ± 7 years, 58% were female and 18 patients (62%) had atrial fibrillation. At baseline, median NT-proBNP was 2440 (1273–3542) ng/l. Etiology of mitral valve disease was classified secondary in 25 patients (85%) and mixed in four patients (15%). According to Carpentieŕs classification, 15 patients (52%) were assigned to Carpentier class I and 14 patients (48%) to Carpentier class IIIb. Mean LVEF was 44 ± 13% (Table 2).
Table 1.
Baseline Patient Characteristics.
| Baseline Data | Overall N = 29 | Non-LVRR N = 19 | LVRR N = 10 | p-Value |
|---|---|---|---|---|
| Clinical characteristics | ||||
| Age (years) | 77 ± 7 | 77 ± 6 | 77 ± 9 | 0.862 |
| BMI (kg/m2) | 25 ± 6 | 24 ± 7 | 26 ± 4 | 0.496 |
| Women, N (%) | 14 (48) | 10 (53) | 4 (40) | 0.699 |
| Hypertension, N (%) | 25 (86) | 15 (79) | 10 (1 0 0) | 0.268 |
| Diabetes mellitus, N (%) | 4 (14) | 2 (11) | 2 (20) | 0.592 |
| Vascular disease, N (%) | 3 (10) | 2 (11) | 1 (10) | 0.999 |
| Coronary artery disease, N (%) | 23 (79) | 15 (79) | 8 (80) | 0.999 |
| Previous CABG, N (%) | 11 (38) | 7 (37) | 4 (40) | 0.999 |
| Previous VS, N (%) | 4 (14) | 3 (16) | 1 (10) | 0.999 |
| Atrial fibrillation, N (%) | 18 (62) | 13 (68) | 5 (50) | 0.432 |
| Logistic EuroSCORE (%) | 27 ± 15 | 28 ± 14 | 26 ± 16 | 0.505 |
| NYHA III/IV, N (%) | 21 (72) | 12 (63) | 9 (90) | 0.201 |
| Labratory assessment | ||||
| Serum Creatinine (mg/dl) | 1.4 ± 0.5 | 1.3 ± 0.4 | 1.5 ± 0.8 | 0.353 |
| Estimated GFR (ml/min/m2) | 49 ± 17 | 49 ± 13 | 49 ± 23 | 0.959 |
| Hemoglobine (mg/dl) | 12 ± 2 | 12 ± 2 | 12 ± 2 | 0.555 |
| NT-proBNP (pg/ml) | 2440 (1273–3542) | 2711 (1388–4408) | 2238 (978–3241) | 0.346 |
Abbreviations: LVRR = Left ventricular reverse remodeling; BMI = Body mass index; CABG = Coronary artery bypass grafting; VS = Valve surgery; NYHA = New York Heart Classification; GFR = Glomerular filtration rate; NT-proBNP = NT-pro-Brain natriuretic peptide.
Table 2.
Baseline CMR parameters according to the presence of left ventricular reverse remodeling.
| CMR at Baseline | Overall N = 29 | Non-LVRR N = 19 | LVRR N = 10 | p-Value |
|---|---|---|---|---|
| LVEDVi (ml/m2) | 103 ± 33 | 105 ± 37 | 99 ± 25 | 0.651 |
| LVESVi (ml/m2) | 60 ± 31 | 68 ± 34 | 46 ± 18 | 0.071 |
| LVSVi (ml/m2) | 43 ± 11 | 38 ± 6 | 54 ± 12 | <0.001 |
| LVEF (%) | 44 ± 13 | 39 ± 12 | 52 ± 10 | 0.002 |
| Effective LVSVi (ml/m2) | 28 ± 8 | 25 ± 7 | 32 ± 8 | 0.082 |
| Cardiac Index (L/min/m2) | 2.0 ± 0.5 | 1.8 ± 0.5 | 2.2 ± 0.5 | 0.060 |
| MR Vol (ml) | 28 ± 17 | 26 ± 18 | 32 ± 16 | 0.038 |
| MR Fraction (%) | 33 ± 16 | 33 ± 18 | 32 ± 11 | 0.651 |
| RVEDVi (ml/m2) | 79 ± 19 | 76 ± 19 | 84 ± 19 | 0.386 |
| RVESVi (ml/m2) | 41 ± 13 | 42 ± 16 | 40 ± 8 | 0.714 |
| RVSVi (ml/m2) | 36 ± 12 | 34 ± 11 | 41 ± 12 | 0.147 |
| RVEF (%) | 46 ± 11 | 45 ± 12 | 48 ± 11 | 0.557 |
| TR Vol (ml) | 16 ± 15 | 15 ± 15 | 19 ± 18 | 0.661 |
| TR Fraction (%) | 19 ± 18 | 19 ± 19 | 20 ± 16 | 0.864 |
| GLS (%) | −15 ± 6 | −13 ± 5 | −18 ± 7 | 0.046 |
| Circumferential Strain (%) | –22 ± 12 | −18 ± 8 | −30 ± 13 | 0.013 |
| Radial Strain (%) | 29 ± 17 | 24 ± 13 | 37 ± 20 | 0.071 |
| Presence of LGE, N (%) | 13 (48) | 10 (53) | 3 (30) | 0.433 |
Abbreviations: LVRR = Left ventricular reverse remodeling; LVEDVi = Left ventricular end-diastolic volume index; LVESVi = Left ventricular end-systolic volume index; LVSVi = Left ventricular stroke volume index; LVEF = Left ventricular ejection fraction; MR Vol = Mitral regurgitation volume; RVEDVi = Right ventricular end-diastolic volume index; RVESVi = Right ventricular end-systolic volume index; RVSVi = Right ventricular stroke volume index; RVEF = Right ventricular ejection fraction; TR Vol = Tricuspid regurgitation volume; GLS = Global longitudinal strain; LGE = Late gadolinium enhancement.
One third of patients (34%) displayed LVRR (decrease of LVEDVi > 15%) at follow-up CMR, while the remaining patients (66%) did not (Fig. 1). Table 2 compares baseline CMR parameters of LVRR and Non-LVRR patients. In logistic regression analysis, baseline LVSVi, LVEF, LV circumferential strain and MR volume at baseline were predictors of LVRR at follow-up (Table 3). ROC analysis indicated that LVRR at follow-up was best predicted by baseline LVSVi > 45 ml/m2 (AUC 0.900; 95% CI (0.732–1.000), LVEF > 40% (AUC 0.834; 95% CI (0.693–0.981), LV circumferential strain < -24% (AUC 0.778; 95% CI (0.561–0.995) and MR volume > 27 ml (AUC 0.765; 95% CI 0.544–0.977).
Fig. 1.
(A) CMR images of a patient prior to and post MitraClip implantation. End-diastolic images in a four-chamber view prior to and six months after MitraClip procedure. Post procedure the MitraClip causes a dark distinction artefact. (B) According to the definition of left ventricular reverse remodeling (LVRR) (defined by a decrease of LVEDVi > 15%), one third of patients (34%) experienced LVRR, while the remaining two thirds (66%) did not. Abbreviations: LVEDVi = Left ventricular end-diastolic volume index; LVRR = Left ventricular reverse remodeling.
Table 3.
Prediction of left ventricular reverse remodeling at follow-up by pre-interventional CMR.
| Predictors of LVRR Reverse Remodeling at Follow-up | |||
|---|---|---|---|
| Baseline CMR Parameters | OR | 95% CI | P-Value |
| LVESVi (per ml/m2 increase) | 0.971 | 0.933 to 1.000 | 0.088 |
| LVSVi (per ml/m2 increase) | 1.244 | 1.095 to 1.540 | 0.008 |
| LVEF (per % increase) | 1.152 | 1.051 to 1.314 | 0.010 |
| Effective LVSVi (per ml/m2 increase) | 1.123 | 0.995 to 1.310 | 0.086 |
| Cardiac Index (per l/min/m2) | 1.002 | 1.000 to 1.005 | 0.083 |
| Global Longitudinal Strain (per % increase) | 0.856 | 0.708 to 0.993 | 0.063 |
| Circumferential Strain (per % increase) | 0.892 | 0.789 to 0.976 | 0.031 |
| Radial Strain (per % increase) | 1.054 | 0.998 to 1.113 | 0.091 |
| MR Volume (per ml increase) | 1.099 | 1.022 to 1.212 | 0.027 |
Abbreviations: LVRR = Left ventricular reverse remodeling; LVEDVi = Left ventricular end-diastolic volume index; LVESVi = Left ventricular end-systolic volume index; LVSVi = Left ventricular stroke volume index; LVEF = Left ventricular ejection fraction; MR Vol = Mitral regurgitation volume; RVEDVi = Right ventricular end-diastolic volume index; RVESVi = Right ventricular end-systolic volume index; RVSVi = Right ventricular stroke volume index; RVEF = Right ventricular ejection fraction; TR Vol = Tricuspid regurgitation volume; GLS = Global longitudinal strain; LGE = Late gadolinium enhancement.
One third of patients (30%) with LVRR had evidence of LGE at baseline CMR, while every second patient (53%) in the non-LVRR group showed LGE at initial CMR (p = 0.433)(Table 2). When stratifying patients according to the presence of coronary artery disease (CAD), in the LVRR group all patients with evidence of LGE had CAD, while in the non-LVRR cohort 80% of patients with evidence of LGE had CAD.
In the entire population, MitraClip implantation effectively decreased MR volume from 28 ± 17 ml to 9 ± 17 ml (p < 0.001). Similarly, MR fraction was reduced from 33 ± 16% to 10 ± 19% (p < 0.001). Mean mitral valve gradient after TMVR as assessed by echocardiography was 3.3 ± 1.1 mmHg. At follow-up CMR, we detected a decrease of LVSVi (Fig. 2). LVEDVi decreased numerically without reaching statistical significance. There were no changes in LVESVi and LVEF (Fig. 2). Furthermore, we detected a trend towards improved effective LVSVi (LV forward flow) (Fig. 2). In line with this, there was an increase in cardiac index at comparable heart rates (73 ± 16 beats/min vs. 74 ± 15 beats/minute; p = 0.813)(Fig. 2). Increased effective LVSVi (LV forward flow) following MitraClip implantation correlated with the decrease of MR volume and MR fraction through TMVR (Fig. 3). LV strain analysis demonstrated no changes of LV global longitudinal strain (-15 ± 6% vs. −14 ± 7%; p = 0.548), circumferential strain (–22 ± 11% vs. −20 ± 9%; p = 0.550) and radial strain (29 ± 17% vs. 27 ± 12%; p = 0.701) after MitraClip implantation.
Fig. 2.
Left ventricular volumes and function prior to and 6 months after MitraClip implantation. Abbreviations: LVEDVi = Left ventricular end-diastolic volume index; LVESVi = Left ventricular end-systolic volume index; LVSVi = Left ventricular stroke volume index; effective LVSVi = Effective left ventricular stroke volume index (=LV forward flow); LVEF = Left ventricular ejection fraction.
Fig. 3.
Correlation between changes in (A) mitral regurgitation (MR) volume and (B) MR fraction and change in effective left ventricular stroke volume index (effective LVSVi) at follow-up. Black dots represent data for each individual patient. The figure shows regression line with 95% confidence interval. Abbreviations: See Fig. 2.
By definition, patients with LVRR were characterized by reduced LVEDVi and LVSVi (Table 4). LVESVi, LVEF and LV strain parameters did not change at follow-up examination (Table 4). Moreover, effective LVSVi (LV forward flow) and cardiac index increase numerically without reaching statistical significance at follow-up in LVRR patients (Table 4). Also, in patients without LVRR, MitraClip implantation reduced MR severity (Table 4). As expected, there were no changes of LV volumes and function in this group (Table 4). Furthermore, effective LVSVi (LV forward flow) and cardiac index tended to improve in this group either (Table 4). In any group, we did not detect any reduction in RV volumes nor an improvement of RV function following MitraClip (Table 4).
Table 4.
CMR parameters at baseline and at follow-up in patients with left ventricular reverse remodeling (LVRR) and without (Non-LVRR).
| Non-LVRR N = 19 |
LVRR N = 10 |
|||||
|---|---|---|---|---|---|---|
| CMR Parameter | Baseline | Follow-up | p-Value | Baseline | Follow-up | p-Value |
| LVEDVi (ml/m2) | 105 ± 37 | 103 ± 30 | 0.831 | 99 ± 25 | 74 ± 19 | 0.024 |
| LVESVi (ml/m2) | 68 ± 34 | 67 ± 30 | 0.989 | 46 ± 18 | 35 ± 15 | 0.165 |
| LVSVi (ml/m2) | 38 ± 6 | 36 ± 9 | 0.479 | 54 ± 12 | 40 ± 10 | 0.010 |
| LVEF (%) | 39 ± 12 | 37 ± 12 | 0.619 | 52 ± 10 | 54 ± 13 | 0.982 |
| Effective LVSVi (ml/m2) | 25 ± 7 | 30 ± 7 | 0.143 | 32 ± 8 | 38 ± 8 | 0.101 |
| Cardiac Index (L/min/m2) | 1.8 ± 0.5 | 2.1 ± 0.4 | 0.047 | 2.2 ± 0.5 | 2.7 ± 0.8 | 0.068 |
| MR Vol (ml) | 26 ± 18 | 8 ± 16 | <0.001 | 32 ± 16 | 9 ± 12 | 0.003 |
| MR Fraction (%) | 33 ± 18 | 12 ± 23 | <0.001 | 32 ± 11 | 13 ± 19 | 0.009 |
| RVEDVi (ml/m2) | 76 ± 19 | 78 ± 17 | 0.743 | 84 ± 19 | 89 ± 25 | 0.681 |
| RVESVi (ml/m2) | 42 ± 16 | 43 ± 12 | 0.876 | 40 ± 8 | 42 ± 17 | 0.789 |
| RVSVi (ml/m2) | 34 ± 11 | 35 ± 12 | 0.777 | 41 ± 12 | 47 ± 14 | 0.442 |
| RVEF (%) | 45 ± 12 | 45 ± 11 | 0.909 | 48 ± 11 | 53 ± 21 | 0.327 |
| TR Vol (ml) | 15 ± 15 | 9 ± 6 | 0.795 | 19 ± 18 | 18 ± 7 | 0.951 |
| TR Fraction (%) | 19 ± 19 | 16 ± 11 | 0.961 | 20 ± 16 | 23 ± 19 | 0.785 |
| Global Longitudinal Strain (%) | −13 ± 5 | −13 ± 5 | 0.840 | −18 ± 7 | −16 ± 9 | 0.526 |
| Circumferential Strain (%) | −18 ± 8 | −15 ± 6 | 0.307 | −30 ± 13 | −29 ± 7 | 0.843 |
| Radial Strain (%) | 24 ± 13 | 22 ± 8 | 0.664 | 37 ± 20 | 35 ± 15 | 0.837 |
Abbreviations see Table 2.
During 18 ± 7 months follow-up, 17% of patients died, additional 13% experienced HF-hospitalizations. Seventeen per cent of patients presented with persistent dyspnea according to NYHA class III/IV at follow-up. There were no differences in clinical outcome in patients with and without LVRR at follow-up (Table 5).
Table 5.
One-year clinical outcome in relation to left ventricular remodeling following MitraClip procedure.
| Non-LVRR N = 19 | LVRR N = 10 | p-Value | |
|---|---|---|---|
| All-cause mortality, N (%) | 3 (16) | 2 (20) | 0.775 |
| HF-hospitalization, N (%) | 2 (13) | 1 (13) | 0.999 |
| NYHA class III/IV, N (%) | 2 (13) | 2 (25) | 0.439 |
Abbreviations: LVRR = Left ventricular reverse remodeling; HF-hospitalization = Heart failure hospitalization; NYHA = New York Heart Association.
4. Discussion
The present study investigated changes of ventricular volumes and myocardial function after MitraClip implantation with serial CMR in order to assess determinants of LVRR and their relationship with clinical outcome. The main findings are: 1) One third of patients undergoing MitraClip implantation in our cohort experienced LVRR as defined by a marked reduction of LVEDVi indicating diastolic LV unloading; 2) LVRR at follow-up was associated with baseline LV function and the magnitude of MR at baseline; 3) MitraClip implantation led to an improvement in hemodynamic status illustrated by an increase in effective LVSVi (LV forward flow) and cardiac index. In this regard, stroke volume index better represents improvements in hemodynamic status, rather than ejection fraction or strain parameters; and 4) The increase in effective LVSVi following TMVR was correlated with the decrease of MR volume and MR fraction through MitraClip implantation.
According to the definition of LVRR (decrease of LVEDVi > 15%), in our study one third of patients revealed LVRR. This is a little bit less compared to previous studies that demonstrated LVRR in about 50% of patients undergoing MitraClip procedure, while the definition of LVRR slightly varies [6], [7], [8]. Average decrease of LVEDVi in our cohort was −10 ± 8 ml/m2 which is quite similar with the decrease of LVEDV in patients with secondary MR in the EVERST trial and somewhat less compared to previous studies [6], [8], [12], [13], [14]. Predictors of LVRR were baseline LV function and the magnitude of MR at baseline. Regarding LV function, larger LVSVi, increased LVEF and LV circumferential strain at baseline were associated with LVRR. This is an important finding in patients with secondary MR that might suffer from a potentially irreversible underlying LV pathophysiology (e.g. ischemic cardiomyopathy). In this regard, baseline LVSVi (<45 ml/m2), LVEF (<40%) and LV circumferential strain (<–23%) seem to be capable in identifying subjects that will not experience LV unloading through TMVR or, on the other hand, those patients in whom the increase in afterload counterbalanced the reduction in LV preload [15]. As shown in previous studies, afterload mismatch is associated with lower LVEF and may lead to adverse clinical outcomes [16], [17]. Thus, indication for TMVR needs to be set with caution and further treatment options (e.g. LV assist device implantation) should be discussed early in these patients. Moreover, larger MR volume at baseline was a predictor of LVRR at follow-up. After publication of MITRA-FR and COAPT, we face an ongoing discussion about optimal patient selection for TMVR. In this regard, Grayburn proposed a conceptual framework of “proportionate” and “disproportionate” MR considering magnitude of MR in relation to degree of LV dilatation and function to identify patients that may benefit from TMVR [18]. In the current study, we did not focus on proportionate and disproportionate MR, however, even in our small cohort, MR volume at baseline was larger in the LVRR group compared to the Non-LVRR cohort (Table 2). Thus, the magnitude of MR at baseline seems to be another important factor that is associated with LV remodeling after MitraClip implantation in secondary MR.
However, not only LV function and MR volume, but also pre-existing structural LV alterations are discussed to impact LVRR and may therefore be useful in risk stratification prior mitral valve interventions (e.g. in selecting the appropriate timing for mitral valve surgery in degenerative MR). Van de Heyning et al. included 41 patients with degenerative MR undergoing CMR and showed an association between the presence of LGE and LV remodeling [19]. In our cohort, the presence of LGE was not associated with LVRR at follow-up. Even when stratifying patients to the presence of CAD, there was no association between the presence of LGE and LVRR. Theoretically, patients with potentially irreversible changes of myocardial structure (e.g. ischemic scar tissue) may experience less distinct changes in LV volumes and function. In future trials, T1 Mapping techniques may provide further insights into the impact of myocardial structure on LV remodeling, as T1 Mapping and Extracellular Volume Mapping are more sensitive in detecting not only focal but also diffuse fibrosis with prognostic impact in various patient populations (e.g. in patients with non-ischemic cardiomyopathy) [20].
In the entire cohort, only LVSVi was significantly reduced at follow-up, while there was a slight decrease of LVEDVi (-10 ± 8 ml/m2). LVESVi remained unchanged at second CMR. These findings can be elucidated by a reduction of LV volume overload, and a potentially increase in LV afterload as LVESVi did not decrease [15]. However, this remains speculative because we did not directly measure afterload in this study. LV systolic function (LVEF) as well as strain parameters remained unaffected following TMVR. However, and despite the fact that LVSVi was reduced, effective LVSVi (LV forward flow) and cardiac index rose six months after TMVR. Thus, MitraClip implantation led to a significant decrease of MR volume which further reduced LV preload and resulted in improved hemodynamics following the procedure. In this regard, effective stroke volume index better represents hemodynamic improvements than LVEF or strain parameters, as we observed an increase in effective LVSVi (LV forward flow) while LVEF and strain parameters remained unchanged following MitraClip. However, the improvement in hemodynamic status was not associated with presence of LVRR at follow-up. But rather, the reduction of MR volume and MR fraction through MitraClip implantation were associated with improvements in effective LVSVi (LV forward flow)(Fig. 3). Thus, we assume that the improvement in hemodynamic status is mainly achieved through the reduction of regurgitant flow and thereby diastolic LV unloading. The pathophysiological basis of this phenomenon has been demonstrated by Gaemperli et al. who demonstrated an 21% increase in LV afterload and conversely a 17% decrease of LV preload acutely after MitraClip procedure through invasive measurements [15]. Similarly, in their study the improvement in hemodynamic status was not associated with LV contractility. Likewise, in the study of Lurz et al. a reduction of LV preload following MitraClip implantation did not lead to acutely improved LVEF and LV strain parameters at follow-up CMR 7 days after TMVR [9]. These observations are in accordance with several echocardiographic studies that show an increase in LV forward flow while LVEF remained unaffected after TMVR [5], [6]. The fact that the increase in effective LVSVi following TMVR was correlated with the decrease of MR volume and MR fraction through MitraClip implantation underlines the importance of procedural success with a sufficient reduction of MR. Grayburn et al. investigated LV remodeling during the first year after MitraClip procedure and found that the correction of volume overload was associated with the degree of residual MR [4]. Moreover, in large registries residual moderate to severe MR was associated with adverse prognosis [21], [22]. This together with our findings suggests that a sufficient reduction of MR volume is of utmost importance to achieve hemodynamic improvement, and thus, most clinical benefit for patients.
During one-year follow-up, clinical outcome was similar in patients with and without LVRR. However, we only included a small patient cohort and our study was not powered for outcome analysis. Adamo et al. demonstrated improved clinical outcomes in patients with LVRR in a larger cohort of patients with secondary MR (184 patients) that underwent MitraClip implantation [7]. Similarly, Nita et al. showed in a mixed cohort including 164 patients with primary and secondary MR that LVRR determined by echocardiography was associated with reduced rates of major adverse cardiovascular events two years after MitraClip [23]. Thus, there is evidence for an association between LVRR and improved clinical outcomes.
5. Study limitations
We included a small but distinct patient cohort. Thus, multivariable analysis on predictors of LV remodeling was not performed. However, this is the largest cohort of MitraClip patients published so far, undergoing sequential CMR imaging. We focused on patients with secondary MR only, thus creating a decisive patient population for comprehensive CMR assessment. In this regard, our study population reflects a real-world setting with a typical mixture of inoperable, high- and intermediate risk patients with secondary MR that currently undergo MitraClip implantation. Due to the sample size, cut-off values for prediction of LV remodeling at follow-up cannot be transmitted one-to-one in clinical practice but might be a benchmark.
6. Conclusions
TMVR with the MitraClip device induces LVRR in more than one third of patients. This is accompanied by an improvement in hemodynamic status six months after the procedure, which can be assessed by effective stroke volume index (effective LVSVi), rather than ejection fraction or strain parameters. LV function and magnitude of MR at baseline are predictors of LVRR. The hemodynamic improvement following TMVR is not associated with LVRR at follow-up, but with the reduction of MR through MitraClip implantation. Thus, reduction of regurgitant flow, rather than an increase in LV contractility seems to account for hemodynamic improvement following TMVR.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
None.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.ijcha.2021.100804.
Appendix A. Supplementary material
The following are the Supplementary data to this article:
References
- 1.Feldman T., Kar S., Elmariah S. Randomized Comparison of Percutaneous Repair and Surgery for Mitral Regurgitation 5-Year Results of EVEREST II. J Am Coll Cardiol. 2015;66(25):2844–2854. doi: 10.1016/j.jacc.2015.10.018. [DOI] [PubMed] [Google Scholar]
- 2.Maisano F., Franzen O., Baldus S. Percutaneous Mitral Valve Interventions in the Real World. J Am Coll Cardiol. 2013;62(12):1052–1061. doi: 10.1016/j.jacc.2013.02.094. [DOI] [PubMed] [Google Scholar]
- 3.Kalbacher D, Schäfer U, v. Bardeleben RS et al. (2019) Long-term outcome, survival and predictors of mortality after MitraClip therapy: Results from the German Transcatheter Mitral Valve Interventions (TRAMI) registry. Int J Cardiol 277:35–41. [DOI] [PubMed]
- 4.Grayburn P., Foster E., Sangli C. Relationship between the magnitude of reduction in mitral regurgitation severity and left ventricular and left atrial reverse remodeling after mitraclip therapy. Circulation. 2013;128(15):1667–1674. doi: 10.1161/CIRCULATIONAHA.112.001039. [DOI] [PubMed] [Google Scholar]
- 5.Kamperidis V., Van Wijngaarden SE, Van Rosendael PJ. Mitral valve repair for secondary mitral regurgitation in non-ischaemic dilated cardiomyopathy is associated with left ventricular reverse remodelling and increase of forward flow. Eur Heart J Cardiovasc Imaging. 2017;19(2):208–215. doi: 10.1093/ehjci/jex011. [DOI] [PubMed] [Google Scholar]
- 6.Brouwer H., Den Heijer M.C., Paelinck B.P. Left ventricular remodelling patterns after MitraClip implantation in patients with severe mitral valve regurgitation:mechanistic insights and prognostic implications. Eur Heart J Cardiovasc Imaging. 2019;20(3):307–313. doi: 10.1093/ehjci/jey088. [DOI] [PubMed] [Google Scholar]
- 7.Adamo M., Godino C., Giannini C. Left ventricular reverse remodelling predicts long-term outcomes in patients with functional mitral regurgitation undergoing MitraClip therapy: results from a multicentre registry. Eur J Heart Fail. 2019;21(2):196–204. doi: 10.1002/ejhf.1343. [DOI] [PubMed] [Google Scholar]
- 8.Cimino S., Maestrini V., Cantisani D. 2D/3D echocardiographic determinants of left ventricular reverse remodelling after MitraClip implantation. Eur Heart J Cardiovasc Imaging. 2019;20(5):558–564. doi: 10.1093/ehjci/jey157. [DOI] [PubMed] [Google Scholar]
- 9.Lurz P., Serpytis R., Blazek S. Assessment of acute changes in ventricular volumes, function, and strain after interventional edge-to-edge repair ofmitral regurgitation using cardiacmagnetic resonance imaging. Eur Heart J Cardiovasc Imaging. 2015;16(12):1399–1404. doi: 10.1093/ehjci/jev115. [DOI] [PubMed] [Google Scholar]
- 10.Radunski U., Franzen O., Barmeyer A. Cardiac remodeling following percutaneous mitral valve repair - initial results assessed by cardiovascular magnetic resonance imaging. Rofo. 2014;186(10):951–958. doi: 10.1055/s-0034-1366216. [DOI] [PubMed] [Google Scholar]
- 11.Westenberg J., Van Der Geest R.J., Lamb H.J. MRI to Evaluate Left Atrial and Ventricular Reverse Remodeling After Restrictive Mitral Annuloplasty in Dilated Cardiomyopathy. Circulation. 2005;30(112):437–442. doi: 10.1161/CIRCULATIONAHA.104.525659. [DOI] [PubMed] [Google Scholar]
- 12.Ailawadi G., Lim D.S., Mack M.J. One-Year Outcomes after MitraClip for Functional Mitral Regurgitation. Circulation. 2019;139(1):37–47. doi: 10.1161/CIRCULATIONAHA.117.031733. [DOI] [PubMed] [Google Scholar]
- 13.Giannini C., Petronio A.S., De Carlo M. Integrated reverse left and right ventricular remodelling after MitraClip implantation in functional mitral regurgitation: An echocardiographic study. Eur Heart J Cardiovasc Imaging. 2014;15(1):95–103. doi: 10.1093/ehjci/jet141. [DOI] [PubMed] [Google Scholar]
- 14.Foster E., Kwan D., Feldman T. Percutaneous mitral valve repair in the initial EVEREST cohort: Evidence of reverse left ventricular remodeling. Circ Cardiovasc Imaging. 2013;6(4):522–530. doi: 10.1161/CIRCIMAGING.112.000098. [DOI] [PubMed] [Google Scholar]
- 15.Gaemperli O., Biaggi P., Gugelmann R. Real-time left ventricular pressure-volume loops during percutaneous mitral valve repair with the mitraclip system. Circulation. 2013;127(9):1018–1027. doi: 10.1161/CIRCULATIONAHA.112.135061. [DOI] [PubMed] [Google Scholar]
- 16.Melisurgo G., Ajello S., Pappalardo F. Afterload mismatch after MitraClip insertion for functional mitral regurgitation. Am J Cardiol. 2014;113(11):1844–1850. doi: 10.1016/j.amjcard.2014.03.015. [DOI] [PubMed] [Google Scholar]
- 17.Jogani S., Van de Heyning C., Paelinck B. Afterload Mismatch After MitraClip Implantation: Intraoperative Assessment and Prognostic Implications. J Invasive Cardiol. 2020;32(3):88–93. doi: 10.25270/jic/19.00336. [DOI] [PubMed] [Google Scholar]
- 18.Grayburn P., Sannino A., Packer M. Proportionate and Disproportionate Functional Mitral Regurgitation. JACC Cardiovasc Imaging. 2019;12(2):353–362. doi: 10.1016/j.jcmg.2018.11.006. [DOI] [PubMed] [Google Scholar]
- 19.Van De Heyning C., Magne J., Piérard L.A. Late gadolinium enhancement CMR in primary mitral regurgitation. Eur J Clin Invest. 2014;44(9):840–847. doi: 10.1111/eci.12306. [DOI] [PubMed] [Google Scholar]
- 20.Puntmann V., Carr-White G., Jabbour A. T1-Mapping and Outcome in Nonischemic Cardiomyopathy. JACC Cardiovasc Imaging. 2016;9(1):40–50. doi: 10.1016/j.jcmg.2015.12.001. [DOI] [PubMed] [Google Scholar]
- 21.Puls M., Lubos E., Boekstegers P. One-year outcomes and predictors of mortality after MitraClip therapy in contemporary clinical practice: Results from the German transcatheter mitral valve interventions registry. Eur Heart J. 2016;37(8):703–712. doi: 10.1093/eurheartj/ehv627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Adamo M., Grasso C., Capodanno D. Five-year clinical outcomes after percutaneous edge-to-edge mitral valve repair: Insights from the multicenter GRASP-IT registry. Am Heart J. 2019;217:32–41. doi: 10.1016/j.ahj.2019.06.015. [DOI] [PubMed] [Google Scholar]
- 23.Nita N., Scharnbeck D., Schneider L.M. Predictors of left ventricular reverse remodeling after percutaneous therapy for mitral regurgitation with the MitraClip system. Catheter Cardiovasc Interv. 2020;96(3):687–697. doi: 10.1002/ccd.28779. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.