Prognostic value of right ventricular longitudinal strain in patients with secondary mitral regurgitation undergoing transcatheter edge-to-edge mitral valve repair

Abstract

Aims

To evaluate the prognostic impact of pre-procedural right ventricular longitudinal strain (RVLS) in patients with secondary mitral regurgitation (SMR) undergoing transcatheter edge-to-edge repair (TEER) in comparison with conventional echocardiographic parameters of RV function.

Methods and results

This is a retrospective study including 142 patients with SMR undergoing TEER at two Italian centres. At 1-year follow-up 45 patients reached the composite endpoint of all-cause death or heart failure hospitalization. The best cut-off value of RV free-wall longitudinal strain (RVFWLS) to predict outcome was −18% [sensitivity 72%, specificity of 71%, area under curve (AUC) 0.78, P < 0.001], whereas the best cut-off value of RV global longitudinal strain (RVGLS) was −15% (sensitivity 56%, specificity 76%, AUC 0.69, P < 0.001). Prognostic performance was suboptimal for tricuspid annular plane systolic excursion, Doppler tissue imaging-derived tricuspid lateral annular systolic velocity and fractional area change (FAC). Cumulative survival free from events was lower in patients with RVFWLS ≥ −18% vs. RVFWLS < −18% (44.0% vs. 85.4%; < 0.001) as well as in patients with RVGLS ≥ −15% vs. RVGLS < −15% (54.9% vs. 81.7%; P < 0.001). At multivariable analysis FAC, RVGLS and RVFWLS were independent predictors of events. The identified cut-off of RVFWLS and RVGLS both resulted independently associated with outcomes.

Conclusion

RVLS is a useful and reliable tool to identify patients with SMR undergoing TEER at high risk of mortality and HF hospitalization, on top of other clinical and echocardiographic parameters, with RVFWLS offering the best prognostic performance.

Graphical Abstract

AUC, area under the curve; CI, confidence interval; TAPSE, tricuspid annular plane excursion; S’TDI, Doppler tissue imaging-derived tricuspid lateral annular systolic velocity; RVFWLS, right ventricle free-wall longitudinal strain; RVGLS, right ventricle global longitudinal strain; FAC, fractional area change.

Open in new tabDownload slide

Introduction

Secondary mitral regurgitation (SMR) is common in patients with heart failure (HF) and reduced left ventricular ejection fraction (LVEF) and is associated with poor outcomes, irrespective of other known clinical prognostic markers.^1,2 Transcatheter edge-to-edge repair (TEER) is a valid treatment option for selected symptomatic patients with relevant SMR.^3–5 Several echocardiographic parameters have been associated with poor prognosis in this setting of patients. Specifically, a score derived from the Cardiovascular Outcomes Assessment of the MitraClip Percutaneous Therapy for Heart Failure Patients with Functional MR (COAPT) trial database including four echocardiographic variables, namely LVEF, LV end-systolic dimension, right ventricular (RV) systolic pressure, and tricuspid regurgitation (TR) has been recently proposed.^6,7 However, RV dysfunction (RVD) was an exclusion criterion in the COAPT trial. Nevertheless, approximately 46% of patients with SMR present a variable degree of RVD in the real world.⁸ The prognostic significance of RVD in SMR was evaluated in several studies assessing RVD by tricuspid annular plane systolic excursion (TAPSE), Doppler tissue imaging-derived tricuspid lateral annular systolic velocity (S′TDI), fractional area change (FAC) or RV to pulmonary arterial (PA) coupling.^9–14

In the setting of valvular heart disease (VHD) and HF, RV systolic function evaluated by speckle tracking echocardiography (STE) emerged as a predictor of outcome, even superior to conventional echocardiographic parameter such as TAPSE, S′TDI, or FAC^15,16 which are affected by significant limitations due to their dependency on load conditions, angle of insonation, and geometric assumptions.¹⁷ Indeed, RV longitudinal strain (RVLS) can detect subclinical RVD and better predict patient outcome in conditions of pressure or volume overload.¹⁷ Nevertheless, the role of RVLS in patients with SMR undergoing TEER has not yet been evaluated. Thus, the aim of this study was to investigate the prognostic value of pre-procedural RVLS in patients with SMR undergoing TEER in comparison with more conventional echocardiographic parameters of RV systolic function.

Methods

This is a retrospective observational study including prospectively collected patients with symptomatic moderate/severe (3+) or severe (4+) SMR undergoing mitral TEER between January 2019 and March 2022 at two Italian high volume centres (ASST Spedali Civili, Brescia and San Raffaele Hospital, Milan). Poor quality of the echocardiographic images not allowing off-line analysis of RV strain and patient lost to follow-up were exclusion criteria.

Echocardiographic assessment

All the echocardiographic images were recorded on digital media storage at the echocardiographic laboratories of each institution and analysed offline by two expert echocardiographers, blinded to patient’s clinical data and outcome.

Secondary MR and TR were graded according to the European Association of Cardiovascular Imaging (EACVI) recommendations.¹⁸ In particular, the vena contracta width was measured at the narrowest point of the regurgitant jet, and the effective regurgitant orifice area (EROA) was calculated with the proximal isovelocity surface area method.¹⁸ Cardiac chamber quantification and evaluation of systolic and diastolic function were performed according to the most recent international guidelines of the American Society of Echocardiography (ASE) and the EACVI.^19–21 We considered as cut-off for identification of RV systolic dysfunction TAPSE <17 mm, S′TDI < 9.5 cm/s, and FAC <35% and as cut-off of RV dilatation a RV mid diameter >35 mm. Right ventricular myocardial deformation analyses were performed from the RV-focused apical four-chamber view, using AutoStrain RV (TOMTEC-ARENA 2020, TomTec Imaging Systems GmbH, Unterschleissheim, Germany). The RV endocardial border was automatically placed in end-diastole and followed during the entire cardiac cycle. The region of interest was manually adjusted in end-systole and in end-diastole if needed. We calculated the RV global longitudinal strain (RVGLS) by averaging the values of all the six RV segments and the RV free-wall longitudinal strain (RVFWLS) by averaging the values obtained from the three free-wall segments, as recommended in the joint consensus document issued by the ASE/EACVI/Industry Task Force.²²

Study outcomes

The primary study outcome was a composite of all-cause death or HF hospitalization at 1 year follow-up. The secondary outcome was all-cause death. Information concerning survival and hospitalization was obtained via: (i) telephone interview with the patient, or if deceased, with family members; (ii) contact with the patient’s physician(s); and (iii) review of electronic medical records. Cause of death was determined based on a review of death certificates, medical records for patients who died while hospitalized, and contact of patient’s physician(s), and it was classified as cardiovascular or non-cardiovascular death according to the Mitral Valve Academic Research Consortium criteria.²³ In case of discrepancy between the different source of information, electronic medical records were considered. Assignment of clinical events was performed by physicians unaware of the patient’s RV systolic function values.

Statistical analysis

The normal distribution of continuous variables was tested with Kolmogorov–Smirnov test. Continuous variables are reported as mean ± standard deviation or median (interquartile range, IQR) and were compared using the Student’s t-test or Mann–Whitney U test, respectively. Categorical variables are reported as counts and percentages and were compared using the χ² or Fisher’s exact tests, as appropriate. The reproducibility was assessed by intra-class correlation coefficients (ICCs) and concordance using the Bland–Altman analysis. An excellent agreement was defined as ICC >0.80. Time-dependent receiver operating characteristic (ROC) curves were used to determine the optimal RVFWLS and RVGLS cut-off values for the primary outcome. Cox proportional hazard regression analysis was used to investigate the association between clinical and echocardiographic parameters with study outcome. All variables differently distributed at univariate analysis (P < 0.05) and variables judged to be clinically relevant were included in the multivariable models, limiting the number of covariates in order to avoid overfitting. The results are shown as hazard ratio (HR) with corresponding 95% confidence interval (CI). The proportional hazard assumption was verified testing the correlation of Schoenfeld residuals with the event time, which failed to reject the null hypothesis that HR remains constant over time. Variables related to RV systolic function were included in the multivariate model separately to avoid collinearity bias. For each multivariable model, Harrell’s C-index was calculated to evaluate its predictive power. The cumulative incidence of all-cause mortality or HF hospitalization was estimated using the Kaplan–Meier method. A sensitivity analysis was performed stratifying the population according to TR severity (trivial or mild vs. more than mild). All statistical analyses were performed using the SPSS software, version 26 (SPSS Inc., Chicago, IL, USA) and Stata, version 14 (Stata Corp, College Station, TX, USA). A two-sided significance level of P < 0.05 was considered statistically significant.

Results

Among 175 patients with moderate/severe (3+) or severe (4+) SMR undergoing mitral TEER between January 2019 and March 2022, 33 patients (19%) were excluded from the present analysis because of inadequate echocardiographic quality for RV strain (N = 28) or because were lost at follow up (N = 5). RV strain measures were obtained from 142 patients who represented the final study population (Figure 1), 69 from San Raffaele Hospital, Milan and 73 from ASST Spedali Civili, Brescia.

Intra- and inter-observer agreement for RVFWLS and RVGLS was reported in Supplementary data online, Table S1.

Baseline characteristics

Median age was 76 years (IQR: 70–80), and 72% were male. Most patients had a moderate LV dilatation (mean indexed end-diastolic volume 93 ± 33 mL) and a severe reduction of LVEF (median 35%, IQR: 28–45). Eighty-nine patients (63%) had ischaemic SMR. Median value of EROA was 0.33 cm² (IQR: 0.28–0.40) and 113 patients (80%) were classified as severe MR. Less than one third of patients had RVD assessed by conventional echocardiographic parameters; in particular, 26% had a TAPSE <17 mm; 33% had a S′TDI <9.5 cm/s; and 28% had a FAC < 35%. Mean value of RVFWLS was −18.6 ± 5.1%; mean value of RVGLS was −15.5 ± 5.2% (Tables 1 and 2).

Patients with RVFWLS ≥ −18% (N = 60, 42%) had significantly higher median values of N-terminal pro-brain natriuretic peptide (NT-proBNP) (3765 vs. 2168 ng/L, P = 0.005), lower median values of LVEF (32% vs. 38%, P = 0.008), more dilated RV (59% vs. 40%, P = 0.042), lower average values of S′TDI (9.6 vs. 10.7 cm/s, P = 0.003), FAC (33% vs. 40%, P < 0.001), and RVGLS (−11 vs. −18%, P < 0.001), and more frequent significant TR (80% vs. 63%) compared to those with RVFWLS < −18% (Tables 1 and 2).

Complete baseline characteristics of the study population, stratified by RVGLS ≥ or < −15% are presented in Supplementary data online, Table S2.

Procedural results

Procedural success (i.e. residual MR ≤ 2 at discharge) was achieved in 129 patients (91%). One, two, or more clip were implanted in 75 (53%), 61 (43%), and 6 (4%) patients, respectively. No significant differences were observed in the population stratified by RVFWLS ≥ or < −18% (see Supplementary data online, Table S3).

Clinical outcomes

Median follow-up was 378 (IQR 367–396) days. At 1-year follow up, the primary outcome of death for any cause or hospitalization for HF was reached by 45 patients (32%). Specifically, 10 patients died (22%), most of them for cardiovascular causes (N = 9), and 35 (78%) were hospitalized for HF. Patients with events had more frequent history of coronary artery bypass graft, higher values of NT-proBNP, worse renal function and worse values of RVFWLS, RVGLS and FAC (see Supplementary data online, Table S4).

At ROC analysis, performed to initially stratify the population, RVFWLS showed the best performance. Indeed, RVFWLS ≥ −18% was associated with the primary outcome with a sensitivity of 72%, specificity of 71% and an area under curve (AUC) of 0.78 (P < 0.001). Also, RVGLS ≥ −15% was predictive of events with a good performance (sensitivity 56%, specificity 76%, AUC 0.69, P < 0.001) (Figure 2). Otherwise, prognostic performance was suboptimal for TAPSE, S′TDI and FAC (AUC 0.52, P = 0.724; AUC 0.53, P = 0.569 and AUC 0.61, P = 0.065 respectively).

Overall survival free from primary endpoint was 68.3% at 1-year follow-up. Cumulative survival free from events was lower in patients with RVFWLS ≥ −18% vs. RVFWLS < −18% (44.0% vs. 85.4%; Log rank P < 0.001) as well as in patients with RVGLS ≥ −15% vs. RVGLS < −15% (54.9% vs. 81.7%; Log rank P < 0.001) (Figures 3 and 4).

Predictors of clinical outcome

At univariate analysis, baseline glomerular filtration rate estimated with the Chronic Kidney Disease—Epidemiology Collaboration (CKD-EPI) equation and NT-proBNP as well as FAC, RVFWLS and RVGLS were associated with an increased risk of the primary composite outcome (Table 3). At multivariate analysis, only FAC, RVGLS and RVFWLS were found to be independent predictors of outcome. Moreover, the identified cut-off of RVFWLS and RVGLS both resulted independently associated with the primary outcome [HR 5.98 (95% CI 2.61–13.7), P < 0.001; HR 3.18 (95% CI 1.45–6.97), P = 0.004, respectively]. The predictive power estimated with the C-index was higher for the multivariable models including RVFWLS (Table 4). RVLS parameters and FAC were also predictive of all-cause mortality alone (see Supplementary data online, Table S5).

Sensitivity analysis

Clinical and echocardiographic characteristics of the population stratified by TR severity [more than mild (N = 100) vs. trivial or mild (N = 42)] were reported in Supplementary data online, Table S6. RVFWLS and RVGLS were predictors of outcome regardless of the degree of TR, whereas FAC resulted significantly associated with events only in presence of trivial or mild TR (see Supplementary data online, Table S7).

Discussion

Our study shows that RV systolic dysfunction is a strong predictor of outcome in patients with SMR undergoing TEER on top of other clinical and echocardiographic parameters. In particular, FAC, RVFWLS and RVGLS have an independent prognostic value, with RVFWLS offering the best performance as demonstrated by both ROC analysis and Harrel’s C index. On the other hand, conventional echocardiographic parameters such as TAPSE or S′TDI do not have a prognostic value in this setting of patients.

TEER emerged as a safe treatment option for patients with SMR and HF.^4,5,24 The two landmark trials investigating whether TEER was superior to conservative management in this setting showed opposite results^3,25 leading to efforts in improving patient selection for TEER.^26–29 Echocardiographic evidence of moderate to severe RVD was an exclusion criterion in the COAPT, but it was not considered in the Multicenter Randomized Study of Percutaneous Repair with the MitraClip Device for Severe Secondary MR.^3,25

Indeed, RVD may be in this setting of patients a prognostic marker of pivotal importance. It has already been demonstrated that it is a marker of prognosis in HF and in VHD.^4,5 Several observational studies investigated the role of baseline RV function in patients with SMR undergoing TEER mainly evaluating the more conventional echocardiographic parameter or RV function such as TAPSE or S′TDI.^10,30,31 The impairment of RV-PA coupling evaluated with TAPSE/PA systolic pressure (PASP) has recently shown to be a major predictor of adverse outcome in patients with SMR undergoing TEER.^12,13

In recent years, there was an increasing application of STE in the evaluation of RVD in the setting of VHD and HF; RVLS has shown to be a major predictor of outcome already in the early stages of the diseases, even superior to conventional echocardiographic parameters of RVD.^32,33

In the setting of patients with SMR undergoing TEER, RVLS parameters showed high discriminative values in predicting unfavourable surrogate outcome such as persistent symptoms (New York Heart Association class > II) after the procedure.³⁴ Moreover, a post-hoc analysis of the COAPT trial showed that RV-PA coupling, assessed by the ratio between RVFWLS and PASP, was a powerful predictor of death or HF hospitalization in both TEER and control arms.¹⁴

To the best of our knowledge, this is the first study evaluating the prognostic value of RVD evaluated by means of STE in a real-world cohort of patients with SMR undergoing TEER. Among all the parameters assessing RVD, RVFWLS, RVGLS and FAC showed an independent prognostic role, while TAPSE and S′TDI did not.

Current EACVI/ASE recommendations suggest a multi-parametric approach for the evaluation of RV function.²² However, global RV function is due for about 80% to the shortening of RV longitudinal fibres, thus the evaluation of the longitudinal excursion of RV free wall could be considered a good approximation of global RV function.³⁵ TAPSE and S′TDI reflect the spatial movement of lateral tricuspid annulus towards the echocardiographic transducer and so are strongly angle-dependent.³⁶ Moreover, these parameters may be overestimated in patients with pulmonary hypertension and LV compression or severe TR.^35–37 FAC has been shown to provide a more accurate estimate of RV systolic function compared with TAPSE in a large meta-analysis on correlation of these parameters with cardiac magnetic resonance imaging-derived RVEF,³⁸ and our data support this thesis. However, the location of the RV immediately posterior to the sternum, its unusual crescent shape and heavy endocardial trabeculation makes difficult to achieve a precise RV endocardial evaluation and FAC may not always be a correct estimation of RVEF,¹⁹ especially in patients with enlarged RV or volume overload conditions.³⁹ This limitation, in addition to the possible overestimation of RV function by FAC in presence of TR, could explain the lack of a prognostic impact of FAC in patients with moderate or severe TR.

Right ventricular longitudinal strain is a direct measurement of intrinsic RV myocardial deformation along the longitudinal plane, which is less affected by loading conditions and geometric assumptions than the previous discussed parameters. Indeed, the prognostic value of both RVFWLS and RVGLS in our study was independent of the degree of TR. The higher prognostic value of RVFWLS over RVGLS has already been reported in patients with HF,⁴⁰ but it was not yet explored in patients with SMR undergoing TEER. Actually, RVGLS reflects deformation from both free-wall and inter-ventricular septum and may therefore be underestimated by concomitant LV dysfunction.

Although reference range and lower limit of normal RVFWLS and RVGLS are not clearly established in literature and are vendor-specific,⁴¹ we identified by ROC analysis two cut off values strongly associated with outcomes, −18% for RVFWLS and −15% for RVGLS. These cut-off values are similar to those previously found in healthy subjects and in the context of HF.^41,42,15

A variable degree of improvement in RV function was observed early after TEER,^34,43 maybe due to reduction in left-sided filling pressure, left atrial reverse remodelling and reduction in systolic pulmonary pressures and is associated with improved prognosis.^11,31 However, in a non-negligible proportion of patients this favourable cascade of events is not observed,^9,44 probably because of irreversible pre-capillary pulmonary hypertension and severe myocyte contractile impairment.

The result of our study suggests that there might be a point of no-return after which TEER in patients with SMR may be futile (i.e. irreversible RVD). The cut-off identified for RVFWLS and RVGLS (i.e. −18% and −15% respectively) could help to recognize the degree of RVD beyond which a benefit in terms of contractile improvement and event-free survival after TEER is no longer expected, although these data need to be confirmed in larger prospective studies.

Limitations

The main limitation of this study is its retrospective nature; a variable degree of selection bias cannot be excluded. Moreover, the sample size is small, thus a type II error may have affected the results of comparisons between groups. Larger studies are needed to confirm our results. Complete echocardiographic follow-up data are not available; however, change in these parameters between baseline and follow-up is beyond the objectives of the study. Data concerning LVGLS and RV 3D volumes and EF were not available for a large number of patients and were not included in the analysis.

Conclusions

Right ventricular systolic function evaluated with STE is a useful and reliable tool to identify patients with SMR undergoing TEER at high risk of mortality and HF hospitalization, with RVFWLS offering the best prognostic performance. Future larger studies are needed to confirm the role of this parameter and allow a better selection of patients who can sustainably benefit from TEER.

Supplementary data

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

Funding

None declared.

Data availability

The data underlying this article are available in the article and in its online supplementary material (see Supplementary data online, Tables S1, S2, S3, S4, S5, S6, and S7).

References

Author notes