Prognostic value of mitral valve haemodynamic parameters obtained by intraprocedural echocardiography in transcatheter edge-to-edge repair

Abstract

Aims

To assess whether intraprocedural transesophageal echocardiographic (TEE)-derived haemodynamic parameters predict outcomes in patients undergoing transcatheter edge-to-edge repair (TEER) for mitral regurgitation (MR).

Methods and results

This is a single-centre, retrospective analysis encompassing 458 (IQR, 104–1035) days of follow-up after 926 consecutive patients [481 (52%) with functional MR] referred to an isolated mitral TEER between 2013 and 2020. Cases without actual clip deployment, or in whom prior mitral procedures had taken place, were excluded. The primary outcome was the combined rate of all-cause mortality or heart failure (HF) hospitalizations. Secondary endpoints included single components of the primary outcome, as well as MR severity at one month and one year following the procedure. A multivariable analysis identified two intraprocedural echocardiographic observations made after clip deployment as independent predictors of the primary outcome: an above mild MR (HR for whole study period 1.49, 95% CI 1.05–2.13, P = 0.026) and a 100% or more increase from baseline in the transmitral mean pressure gradient (TMPG) (HR for whole study period 1.32, 95% CI 1.01–1.72, P = 0.039). Also, MR grade of above mild and the absence of a normal pulmonary venous flow pattern (PVFP) bilaterally were associated with an increased risk for HF hospitalizations and greater-than-mild 1-month MR. No prognostic role was demonstrated for the change in MR severity, the absolute TMPG, or the mere improvement in PVFP.

Conclusion

Immediate post-TEER MR severity and the relative change in TMPG are predictive of clinical and echocardiographic outcomes following the procedure.

Graphical Abstract

Prognostic value of immediate post-clipping haemodynamic parameters obtained by intraprocedural echocardiography in mitral transcatheter edge-to-edge repair. Analysing the data of 926 consecutive mitral transcatheter edge-to-edge procedures performed between 2013 and 2020, several post-clipping haemodynamic variables observed on intraprocedural transesophageal echocardiogram were identified that predicted clinical and echocardiographic outcomes. Of them, above mild residual regurgitation and a 100% or greater rise in transmitral mean pressure gradient were associated with a higher risk for the combined endpoint of all-cause mortality or heart failure hospitalizations at 1-year and along the entire follow-up period. The former, as well as pulmonary venous flow pattern non-normalization (i.e. an S/D velocities ratio of <1 on both sides), also predicted 1-year heart failure hospitalizations and 1-month regurgitation severity above mild. E/A ratio, available for 50% of the cohort, was also predictive of the combined outcome at 1-year. No prognostic role was demonstrated for other echocardiographic parameters explored. Numbers in the cells denote hazard or odds ratios, as appropriate. HF = heart failure; MR = mitral regurgitation; PHT = pressure half-time; PVFP = pulmonary venous flow pattern; TMPG = transmitral mean pressure gradient.

Open in new tabDownload slide

Introduction

Transesophageal echocardiography (TEE) has transformed structural heart interventions as it enables a real-time, high-resolution delineation of cardiovascular structures and demonstration of hemodynamics, all while avoiding exposure to radiation and iodinated contrast media. In mitral transcatheter edge-to-edge repair (TEER), the utilization of TEE is invaluable in grading and classifying valvular pathology, as well as in guiding various stages of the procedure and evaluating the end result.^1,2 Recently, valve-related observations made by TEE immediately after clip deployment have been shown to possess prognostic implications as well, potentially allowing for real-time device optimization. Specifically, a more significant residual mitral regurgitation (MR), a higher transmitral mean pressure gradient (TMPG), and the persistence of pulmonary venous flow pattern (PVFP) aberrations have all been correlated with adverse outcomes.^3–7 However, this was based on either small-scale studies, cohorts that included cases without actual clip deployment or who were exposed to non-TEER interventions, or analyses that were adjusted for non-procedural variables. Using a large, contemporary database of patients undergoing an isolated, first-ever, technically successful mitral TEER, along with focusing on intraprocedural echocardiographic parameters only, we aimed to verify the results of these prior studies and comprehensively characterize the association between haemodynamic TEE measurements obtained immediately after clip deployment and downstream clinical and echocardiographic outcomes following the intervention.

Methods

Study population and outcomes

Our study is based on the Cedars-Sinai Medical Center (CSMC) registry of consecutive mitral TEER procedures performed on adult patients between 1 January 2013 and 31 December 2020. We excluded cases subjected to concurrent non-TEER interventions, as well as those who had previously undergone mitral procedures. Technical failure (i.e. the absence of clip deployment) also disqualified patients from analysis.

The primary outcome was defined as the composite of death from any cause or hospitalizations for heart failure (HF) along the study course. Secondary outcomes included separate components of the primary outcome and residual MR severity that is more than mild after one month and one year.

The Cedars-Sinai Institutional Review Board (IRB) approved the study and waived the need for informed consent due to its observational nature.

Procedural aspects

All interventions were carried out following a Heart Team discussion that included at least one general cardiologist, one interventional cardiologist, and one cardiac surgeon. Procedures utilized the MitraClip^TM system (Abbott Vascular Inc, Santa Clara, CA, USA) and were performed under general anaesthesia, via a femoral venous approach, and with echocardiographic and fluoroscopic guidance. Monitoring by a right heart catheterization (RHC) was employed as well.

Echocardiographic assessment

Intraprocedural echocardiograms were performed by a team of experienced echocardiologists, each recognized by the institutional faculty to have obtained a level III training in echocardiography.⁸ Interpretation was blindly accomplished by two study members (AS and ML), also holding a level III certificate, who were not involved in patient care. The ultrasound system used was EPIQ (Philips, Amsterdam, Netherlands). Post-test processing was performed using the PICOM365 platform (SciImage, Los Altos, CA, USA) and QLAB, version 12.0 (Philips, Andover, MA, USA) for two- and three-dimensional (3D) parameters, respectively. All examinations and reports complied with the relevant American Society of Echocardiography (ASE) guidelines.^9,10 Accordingly, haemodynamic parameters of interest were assessed by multiple standard views that exposed at least one pulmonary vein on either side. In every study, the highest or average values of continuous variables were considered for analysis based on the presence of a sinus or a non-sinus rhythm, respectively. For each variable, the median values of the entire cohort were calculated.

Mitral valve hemodynamics included pre- and post-clip deployment transvalvular pressure gradients and pressure half-time by continuous-wave (CW) Doppler, as well as pulsed-wave (PW)-derived E-wave peak velocity and the ratio of the E to A waves peak velocities, where applicable. PVFP was similarly determined before and after clip placement using the ratio of peak systolic (S) wave velocity to peak diastolic (D) wave velocity as measured by a PW Doppler beam placed within 1 centimeter of the PV ostia in the left atrium. An S/D ratio of 1 and above, between 1 and zero, and below zero denoted normal, blunted, and reversed patterns, respectively. PVFP improvement was defined as any rise in the S/D velocities ratio on either side compared to the baseline, while PVFP normalization required the emergence of an S/D velocity ratio of ≥1 on any side.

MR grade was based on the integration of qualitative and quantitative measures, when possible. To better appreciate the continuous nature of MR severity, regurgitation at all stages was graded on a 0-to-6 scale, in which even values represented ‘complete’ degrees (i.e. 0 = no, 2 = mild, 4 = moderate, and 6 = severe) and odd values denoted ‘in-between’ degrees (where 3 = mild-to-moderate and 5 = moderate-to-severe); a value of 1 was used for trace/minimal MR. The valve area following clip deployment was retrospectively derived using 3D planimetry from reconstructed images.

Pre- and post-procedural transthoracic echocardiographic (TTE) studies followed similar principles. Left ventricular ejection fraction (LVEF) was assessed using Simpson’s biplane methods of disks. Pulmonary arterial systolic pressure (PASP) was calculated by adding the maximal CW-derived tricuspid regurgitation (TR) pressure gradient, as obtained in the right ventricular (RV) inflow view, the parasternal short-axis view, the apical views, or the subcostal views, with the estimated right atrial pressure (RAP). One of three values was assigned to the estimated RAP – 3, 8, or 15 mmHg – based on a normal sized (≤2.1 cm), normally collapsed (>50% at brisk inspiration) inferior vena cava (IVC) on the subcostal views; a dilated, non-normally collapsed IVC; or any other combination, respectively.

Data collection and analysis

Demographic characteristics, medical conditions, electrocardiographic (ECG) and imaging studies results, procedural details, and outcome measures and complications were all extracted from CS-Link^TM (Epic, Verona, WI, USA), a web-based patient medical record portal that incorporates data from local hospitals, outpatient medical providers, and state authorities. The most recent Mitral Valve Academic Research Consortium (MVARC) definitions were used.¹¹

Variables were reported as frequencies and percentages, means and standard deviations (SDs), or medians and interquartile ranges (IQRs), as appropriate. The change in invasively acquired haemodynamic parameters from baseline was evaluated by the Wilcoxon text. Correlation between various echocardiographic deltas utilized the Pearson’s coefficient. Inter-observer variability was assessed using the interclass correlation coefficient (ICC).

To identify echocardiographic predictors of the prespecified endpoints, a univariable Cox regression hazard ratio analysis was employed, after which variables with a P-value of <0.1 were integrated into a multivariable model. Regarding the primary outcome, two such models were constructed in a successive fashion—the first one considering all available intraprocedural parameters, and the second one including prognostically significant variables revealed by the former and baseline (i.e. TTE-derived) LVEF and PASP. Predictors for MR severity at 1-month and 1-year and determinants of the change in intraprocedural TMPG were explored using binary logistic regression models. Further characterization of these determinants was made by Harrell’s C and Youden’s indices. The risk for the occurrence of clinical outcomes was graphically displayed according to the Kaplan-Meier method, with comparisons of cumulative event-free survival across strata by the Log-Rank test.

Cases with missing values were censored from the relevant analyses. A two-sided P-value of <0.05 defined statistical significance. All calculations were performed using SPSS^TM Statistic for Windows software, version 24 (IBM Corporation, Armonk, NY, USA).

Results

Baseline characteristics of the study population

Overall, 926 patients met inclusion criteria and were followed for a median of 458 (IQR, 104–1035) days (Figure 1). One-month and one-year post-TEER echocardiograms were performed at a median of 33 (IQR, 29–38) and 370 (IQR, 355–403) days after the intervention, respectively.

The study cohort comprised of mostly (n = 547, 59%) men with a median age of 79 (IQR, 70–86) years who experienced disabling symptoms and a high burden of comorbidities (Table 1), as evident by a New York Heart Association (NYHA) class of III-IV in 93% of patients and a median Kansas City Cardiomyopathy Questionnaire (KCCQ) 12 score of 38 (IQR, 17–60) points. MR was graded as severe in 77% of cases and classified as secondary (functional) in 481 (52%) patients, primary in 395 (43%), and mixed in 50 (5%). Left heart function and dimensions were variable at baseline, with a mean left ventricular (LV) ejection fraction (EF) of 46.9 ± 19.5% and mean LV end-diastolic and end-systolic diameters (EDD and ESD) of 5.4 ± 1.11 and 4.1 ± 1.4 cm, respectively.

Procedural aspects and haemodynamic results

The great majority of patients were treated by 1 or 2 devices, which were mainly of the 1st or 2nd generation (Table 2). Elevated filling pressures observed initially, as expressed in the left atrial V wave and the mean left atrial pressure (LAP), were significantly reduced following clip deployment (P < 0.001).

Intraprocedural transesophageal echocardiographic observations

A detailed description of intraprocedural TEE measurements is presented in Table 3. Inter-observer reliability was good, with ICC values of 0.90–0.93 (all P < 0.001) for all variables. Prior to clip deployment, the median TMPG was 2 (IQR, 1–2) mmHg. Mitral pressure half-time, E wave velocity, and E/A velocities ratio (available for 50% of the study’s cohort) were 60 (IQR, 45–78) msec, 112 (IQR, 93–132) cm/sec, and 1.8 (IQR, 1.3–2.5), respectively. PVFP was abnormal in one or more veins in 93% of studies, and 50% of cases exhibited a complete S/D velocities ratio reversal on at least one side. After TEER, MR was reduced by a median of 4 (IQR, 3–4) grades compared to baseline, to mild or less in 671 (73%) patients. At the same time, the TMPG decreased, remained unchanged, or rose by less than 100% from the pre-clipping phase in half of the visualizable images (n = 414, 50%), as reflected by a median delta of 100% (IQR, 0–150%); the absolute post-TEER value of the TMPG was 3 (IQR, 2–4) mmHg. Mitral pressure half-time became 103 (IQR, 72–139) msec, E wave velocity was 130 (IQR, 110–149) cm/sec, and E/A velocities ratio (again available only in 50% of studies) was 1.24 (IQR, 0.96–1.63). PVFP improved in 78% cases and normalized in 64%. Notably, a direct correlation was observed between MR regression and decrease in the V wave (Pearson’s r = 0.75, P = 0.030) and mean LAP (Pearson’s r = 0.69, P = 0.045). The 3D mitral valve area after clip deployment, available for post-test processing in 573 (61.9%) patients, was 1.69 (IQR, 1.31–2.37) cm².

Primary outcome

A total of 342 (37%) patients died or were hospitalized due to HF during the follow-up period. The corresponding figures at 1-year, median follow-up (i.e. 458 days), and 4 years were 236 (26%), 255 (28%), and 324 (35%), respectively. Whole study period rates of the primary outcome differed substantially between various MR etiologic subgroups: 46% (functional MR), 26% (primary MR), and 32% (mixed MR) (P < 0.001).

After multivariable analysis (Table 4), two post-clip deployment observations made by intraprocedural TEE emerged as independently predictive of the primary outcome at all time points explored: a residual MR severity that is greater than mild (whole period HR 1.49, 95% CI 1.05–2.13, P = 0.026) and an increase in the TMPG which is equal to or higher than the cohort’s median (i.e. a ≥ 100% rise) (whole period HR 1.32, 95% CI 1.01–1.72, P = 0.039). This was also demonstrated upon consideration of baseline LVEF and PASP, both of which were not fully assessed in most intraprocedural studies (see Supplementary data online, Table S1). Accordingly, as the TMPG increased more from baseline, the event-free survival probability further decreased (see Supplementary data online, Figure S1).

Among cases in whom there was a visible A wave (n = 457), a third parameter was identified that predicted the primary outcome at 1-year: E/A velocities ratio as a continuous variable (HR 1.32, 95% CI 1.10–1.58, P = 0.002). No prognostic role was shown for the absolute value of the immediate post-clip deployment TMPG, both as a continuous variable and as a dichotomous one (using three cut-offs – 5, 6, and the cohort’s median of 3 mmHg). Similarly, neither PVFP improvement nor normalization forecasted the combined endpoint.

As presented in a Kaplan-Meier analysis (Figure 2), cumulative event-free survival rate at 4 years after the procedure was 68 ± 5% in patients with ≤ mild MR and 38 ± 5% in those with > mild MR (P < 0.001). Likewise, it was 52 ± 3% in patients exhibiting ΔTMPG below the cohort’s median of 100% and 38 ± 4% among those with higher ΔTMPG values (P = 0.039). By a Cox regression analysis, the risk for the primary outcome at 4 years in patients with one or two of the echocardiographic predictors mentioned above was 1.5 (95% CI, 1.1–1.9) times or 2.2 (95% CI, 1.6–3.2) times higher, respectively, than the one calculated for those with none (all P < 0.001).

Secondary outcomes

During the study period, all-cause mortality affected 211 (23%) patients and HF hospitalizations occurred in 214 (23%). At 1-year, the rates were 13% (n = 116) and 16% (n = 147), respectively. Echocardiographic data, available in 663 (72%) patients at 1-month and in 354 (38%) patients by 1-year, revealed MR of greater than a mild degree in 245 (37%) and 171 (48%) cases, respectively. A graph describing MR severity at different stages of the study is provided in Supplementary data online, Figure S2.

Overall, no significant association was found between any of the immediate post-TEER TEE haemodynamic observations and neither all-cause mortality nor 1-year MR severity. A residual MR of above mild did predict a trend towards an increased death risk (whole period HR 1.51, 95% CI 0.98–2.30, P = 0.056), though. Also, a greater-than-mild residual MR and a lack of PVFP normalization (i.e. an S/D velocities ratio of <1 on both sides) were associated with increased risks for HF hospitalizations and 1-month MR of more than mild degree. These analyses are summarized in Supplementary data online, Tables S2 and S3, and in Figures 3 and 4.

Determinants of transmitral mean pressure gradient increase

According to a binary logistic regression model, a 100% or more rise in the TMPG following clip deployment compared to baseline was associated with an age of 75 and above prior to TEER, a history of myocardial infarction or coronary revascularization, a functional MR aetiology, and lower baseline TMPG values (see Supplementary data online, Table S4). The use of 2 or more clips conferred higher odds as well for this TEE-derived observation, which did not reach statistical significance. The C-statistic of baseline TMPG for a ≥ 100% rise in the TMPG was modest, at 0.58 (95% CI, 0.54–0.61, P < 0.001). Youden’s index, representing the optimized value of the parameter in terms of sensitivity and specificity, was 2.5 mmHg.

Discussion

This study evaluated the prognostic implication of mitral valve-related haemodynamic parameters observed on intraprocedural TEE immediately after MitraClip deployment. Importantly, we focused on widely used, easily obtainable 2-dimensional measures that have been shown to correlate with outcomes and/or ones that are currently used to assess the haemodynamic consequence of mitral valvulopathies. Regarding PVFP, the velocities ratio was chosen, rather than the velocity time integrals (VTIs) ratio, because of its simpler acquisition and comparable predictive performance.⁶

Our results, also depicted in the Graphical Abstract, may be summarized as follows: 1. A greater than mild residual MR, as well as a 100% or more rise in the TMPG compared to baseline, independently conferred a higher risk for the composite endpoint of all-cause mortality or HF hospitalizations at 1-year and throughout the follow-up period of 458 (IQR, 104–1035) days; 2. An above mild MR and a lack of PVFP normalization bilaterally were associated with an increased risk of HF hospitalizations and of 1-month greater-than-mild MR; and 3. No other echo-Doppler observations made after TEER, including the change in MR grade, the absolute TMPG, or the status of PVFP improvement, were prognostically meaningful.

The study’s findings suggest that the change in valvular hemodynamics following mitral TEER, expressed by the change in TMPG, is potentially a more accurate predictor of the procedure’s aftermath compared to a ‘stand-alone,’ post-clip haemodynamic phase, represented by the absolute TMPG. While both parameters illustrate a trade-off between regurgitation and stenosis, it is arguably only the former that conveys the intervention’s impact on this balance. Accordingly, the higher the increase in the TMPG, the greater the extent to which regurgitation reduction is counteracted by stenosis augmentation, and a less favourable outcome is to be expected. Notably, a lower pre-procedural TMPG value was associated with an exaggerated rise in the post-clipping TMPG, however, discrimination was only modest, thus stressing the importance of intraprocedural measurements in risk stratification.

Apart from highlighting the significance of the change in the TMPG, our study refuted the prognostic value of previously acclaimed variables, such as the absolute TMPG^4,12 and the PVFP,⁵ for most endpoints studied. One explanation for this observation is that these parameters are closely related to MR severity, which has been consistently shown to predict prognosis following TEER. Therefore, their perceived predictive ability for clinical events, which may only reflect the regurgitation degree, should be diminished when analysed simultaneously with MR—as was the case in our novel, comprehensive multivariable models. Another reason for the contradicting findings may lie in statistical power, as our project employed one of the largest mitral TEER cohorts to date for this particular purpose.

Our study suggests two clinical implications. First, its observations could help refine the definition of (procedural) success of mitral TEER, which at present is largely based on MR grade and TMPG as depicted on a post-procedural TTE.¹¹ This is important, as TTE may not be available and/or interpretable for all cases, and because MR assessment on TTE after TEER may be particularly challenging—a notion which could explain the discrepancy between the lack of association between the prognostically meaningful TMPG change and the post-TEER MR grade in our cohort. Furthermore, recent studies have reported mixing results regarding the prognostic utility of TTE-derived TMPG,^13–15 and those that did support a risk stratification role for post-procedural TTE, in general, may have been biased by non-TEER influences, all of which did not apply to our selected cohort of isolated, first-time mitral TEER patients.

A more immediate and practical implication of this study is its potential to assist interventional teams in achieving an optimal result in real-time. Although not explicitly explored by us, corrective measures could theoretically be undertaken intra-procedurally that will allow for the constellation of haemodynamic findings associated in our cohort with improved outcomes. The exact nature of these should be determined by further, perhaps prospective research. What can be argued based on our findings, is that MR reduction should not necessarily be mirrored by TMPG increase, as some of our patients experienced either no change or an actual decrease in the TMPG – again emphasizing the importance of the delicate balance between regurgitation and stenosis.

Limitations

First, our study represents a retrospective analysis from one centre which did not employ an external core laboratory. However, the sample size was large in absolute terms, echocardiograms were performed by highly skilled staff utilizing the same ultrasound systems, and the images were retrospectively read by blinded echocardiologists, all reinforcing consistency and reliability nonetheless. Second, missing data may have affected our conclusions, mostly regarding the 1-month and 1-year MR severity endpoints. Third, immediate post-clipping LVEF and PASP were neither systematically reported nor assessable using stored images of the intraprocedural TEEs, as those included partial views or were absent altogether. Notwithstanding their theoretical prognostic significance, they were consequently not considered in the final analysis. We tried to overcome this by employing LVEF and PASP values acquired during baseline TTE in a separate multivariable model, the result of which was in agreement with the TEE-only analysis. Fourth, other than heart rate and filling pressures, we were not able to retrieve data about potential real-time determinants of the transvalvular gradients, such as volume status and systemic blood pressure. Also, valve area by planimetry was available only for about 60% of patients, some of whom exhibited low-quality images. Yet, the parameters used in our study are also the ones used in clinical everyday practice and are considerably easier to obtain and process than 3D variables, particularly in a time-constrained setting. Fifth, the assessment of post-interventional MR relied more often on qualitative measures due to artefacts imposed by the implanted clip(s). However, we demonstrated a good correlation between MR reduction and invasively measured left atrial pressures to decline. Lastly, our findings may not apply to non-MitraClip systems as these were not used in our institution. Further, they do not relate to 3D indices of MR which, although potentially possessing prognostic abilities as well,¹⁶ are yet to gain widespread acceptance due to a more complicated application and interfering device-related artefacts.

Conclusion

In our experience, an MR grade of greater than mild and doubling of the TMPG compared to baseline following mitral TEER, both as observed on intraprocedural TEE immediately after clip deployment, predicted an increased risk for the composite of all-cause mortality or HF hospitalizations. Concurrently, an above mild MR and a lack of PVFP normalization were associated with an elevated risk for HF hospitalizations and 1-month MR severity above mild.

Supplementary data

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

Acknowledgements

The study was supported in part by the California Chapter of the American College of Cardiology through the Save a Heart foundation.

Data availability

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

References

Abbreviations

 
• MR =

  Mitral regurgitation

 
• PG =

  Pressure gradient

 
• PVFP =

  Pulmonary venous flow pattern

 
• TEE =

  Transesophageal echocardiography

 
• TEER =

  Transcatheter edge-to-edge repair

 
• TMPG =

  Transmitral mean pressure gradient

Author notes