1 INTRODUCTION

Atrial functional mitral regurgitation (aFMR) is associated with left atrial (LA) and mitral annular dilatation most often due to atrial fibrillation (AF) or heart failure with preserved ejection fraction (HFpEF). aFMR was found in up to 53% of HFpEF patients.¹ However, the exact prevalence of aFMR in the general population is unknown and seems currently to be underreported.² Characteristics of aFMR are normal LV-function, normal mitral leaflet motion, a more central mitral regurgitation (MR) jet and severe LA dilatation. This is in contrast to ventricular secondary mitral regurgitation (vFMR), where systolic left ventricular (LV) dysfunction, restricted leaflet motion, an increased tethering force and LA dilatation are pathogenetic. However, severity of MR at discharge after acute decompensation in HFpEF was linked to worse clinical outcome implicating a need for therapeutic options.³ Transcatheter mitral valve repair is an emerging therapeutic option in aFMR, but currently no specific guidelines exist.² However, mitral transcatheter edge-to-edge repair (TEER) is effective and safe in the treatment of aFMR and improved heart failure symptoms in long-term follow-up.⁴ Among patients with mitral TEER between 7.5% and 12.8% of the patients showed aFMR in large multicenter registries indicating a relevant patient population.^4-7 Recently the prospective large scale multicenter MITRA-PRO registry revealed that residual MR assessed by the MitraScore is associated with 1-year mortality and rehospitalization in patients with functional MR (FMR) undergoing mitral TEER.⁸ For residual MR quantification the validated and multimodal MitraScore was used, which included both echocardiographic and hemodynamic parameters.^{8, 9} However, little is known about the impact of residual MR on the prognosis of patients with aFMR especially in comparison to vFMR. One might speculate, that patients with HFpEF might tolerate residual MR after mitral TEER for a longer time compared to patients with vFMR. Therefore, we aimed to evaluate the role of residual MR following TEER in patients with aFMR and compare the results to vFMR.

2 METHODS

3 RESULTS

The MITRA-PRO registry included 1491 consecutive patients with degenerative, functional or mixed (degenerative-functional) MR undergoing mitral TEER and residual MR quantification using the multimodal MitraScore. After exclusion of patients with degenerative MR, mixed MR or nonclassifiable FMR 846 patients with FMR were identified and either classified as aFMR or vFMR (Figure 1). In this subanalysis 722 patients with vFMR and 124 patients with aFMR were included (Figure 2A).

Baseline characteristics of aFMR and vFMR groups are shown in Table 1. Patients in the aFMR group were predominantly female (69.4%) compared to vFMR (35.5%). Furthermore, mean age was significantly higher in the vFMR group, while logistic Euroscore was lower compared to aFMR (76.0 ± 9.0 vs. 81.0 ± 5.0 years and 25.0 ± 15.7 vs. 16.4% ± 11.1%, p < 0.001). As expected, diabetes mellitus and coronary artery disease were more frequent in vFMR, while arterial hypertension was found in 91.1% in aFMR and 83.4% in vFMR (Table 1). LVEF was normal in aFMR and significantly reduced in vFMR (58.5 ± 5.3 vs. 35.2% ± 12.8%, p < 0.001), while systolic pulmonary artery pressure (sPAP) was comparable in both groups (51.0 ± 29.9 vs. 47.3 ± 14.2 mmHg, p = 0.66). Right heart function was more impaired in vFMR (Table 1). Functional status assessed by NYHA classification and 6-min walking distance was comparable in both groups (Table 1). However, NTproBNP was markedly elevated in vFMR (4246.5 vs. 2107.0, <0.001) due to the reduction in LVEF.

The MitraScore was evaluated for all patients prior and postmitral TEER. However, the decrease of MitraScore points pre and post mitral TEER was comparable in the aFMR group compared to the vFMR group (vFMR: pre 9.5 ± 2.1 vs. post 2.7 ± 1.9 and MitraScore ≥4: pre 9.0 ± 2.1 vs. post 2.5 ± 1.8). All variables of the MitraScore in both groups pre and postmitral TEER are displayed in Table 2.

Significantly more TEER devices were used in vFMR compared to aFMR and also procedure time was significantly longer in patients with vFMR (Figure 2B, Table 3). Periinterventional complications were rare in both groups with a significant difference in postinterventional renal failure (vFMR 7.8% vs. aFMR 1.6%, p = 0.013). All procedural parameters are listed in Table 3.

1-year clinical follow-up was completed in 695 patients (96.3%) of the vFMR group and 122 patients (98.4%) of the aFMR group. NYHA classification was comparable between aFMR and vFMR at baseline. In both groups, mitral TEER significantly improved NYHA classification at discharge and 12-months follow-up. No significant difference was found between both groups (Figure 2C). Also, MR classification assessed by transthoracic echocardiography showed comparable baseline values for aFMR and vFMR. Mitral TEER improved MR classification for both groups at discharge and 12-months follow-up. However, no significant difference between aFMR and vFMR was found (Figure 2D).

During the index procedure hospital stay, 9 patients of the vFMR group (1.2%) and 2 patients of the aFMR group (1.6%) died. Furthermore, the in-hospital MACE and MACCE rates were comparable between both groups (Table 3). 1-year mortality was significantly lower in patients with aFMR. Therefore, the mortality accounts for 7.5% in the aFMR group and 18.2% in the vFMR group (Figure 2E).

Patients with aFMR and vFMR were further allocated to MitraScore ≤3 and MitraScore ≥ 4. In the aFMR group, 94 patients showed a postprocedural MitraScore ≤3, while 30 patients had a MitraScore ≥4. Furthermore, in vFMR, 518 patients revealed a mild or no residual MR (MitraScore ≤3) and 204 patients showed a relevant residual MR (MitraScore ≥4) postmitral TEER. Baseline and procedural characteristics of the subgroups are shown in Table 4. In both subgroups 1-year mortality was significantly higher in the MitraScore ≥4 group (Figure 3A). However, 1-year mortality was significantly lower in the aFMR subgroups (aFMR: MitraScore ≤ 3 6.6% and MitraScore ≥ 4 10.3% vs. vFMR: MitraScore ≤ 3 16.1% and MitraScore ≥ 4 23.7%; Figure 3A). No significant difference was found for 1-year rehospitalization (aFMR: MitraScore ≤ 3 29.1% and MitraScore ≥ 4 46.2% vs. vFMR: MitraScore ≤ 3 38.8% and MitraScore ≥ 4 43.2%), while the combined end-point 1-year mortality and reintervention was significantly different between the groups (Figure 3B).

Residual MR (MITRA-Score ≥ 4) was an independent predictor of survival in a multivariate analysis in all FMR patients (n = 846) with a HR (95%): 1.57 (1.09–2.25). Other independent predictors of 1-year mortality were impaired renal function (eGFR < 30 mL/min/1.73 m²) and LVEF. Furthermore, we found a MITRA-Score ≥ 4 to be an independent predictor of 1-year mortality in vFMR (n = 722) with a HR (95%): 1.58 (1.09–2.30). Impaired renal function (eGFR < 30 mL/min/1.73 m²) was also associated with 1-year survival with a HR (95%): 1.94 (1.31–2.88) in vFMR. In aFMR, where the number of patients (n = 122) and the event rate (n = 9) were low, residual MR was statistically no independent predictor of mortality with a HR (95%): 0.89 (0.19–4.22). Of note, no parameter of the multivariate analysis in aFMR was found to be an independent predictor of mortality indicating a low statistical power due to the low event rate in this cohort.

4 DISCUSSION

This analysis of the large-scale multicenter MITRA-PRO registry describes for the first time the impact of residual MR on 1-year survival and rehospitalization in aFMR versus vFMR. Recently feasibility and efficacy of mitral TEER in the context of aFMR were demonstrated by an analysis of the EuroSMR registry.⁴ These effects and clinical improvement were found to be comparable to patients with vFMR.⁴ The prevalence of aFMR was 7.8% in this highly selected TEER cohort. In our analysis we found a comparable prevalence of 8.3%, which is also in line with other studies like the EXPAND trial (12.8%), MITRA-TUNE registry (7.5%) or GIOTTO (6.0%).^{5, 7, 11} Patients with aFMR share typical characteristics compared to vFMR: (1) more female patients, (2) higher mean age, (3) more arterial hypertension and (4) less coronary artery disease. These findings are related to the underlying HFpEF in patients with aFMR. TEER in aFMR showed a good safety profile with a low in-hospital MACCE rate of 2.5%, which is in line with previous publications such as GIOTTO registry (complication rate 2%), MITRA-TUNE (in hospital death 2%) or a Spanish registry (in-hospital death 2.1%).^{7, 11, 12}

In the MITRA-TUNE registry NYHA stage III/IV was found in 35/65% at baseline and 0/10% at follow-up and in the EXPAND trial in 11.3/73.6% at baseline and also 0/21.1% at follow-up of the aFMR patients.^{5, 7} In the MITRA-PRO registry we found NYHA stage III/IV in 13.8/76.4% at baseline and in 4.2/48.4% at 12-months follow-up, which is significantly higher than reported for aFMR. Of note, in the MITRA-PRO registry follow-up clinical assessment was done investigator-independent from a central side to exclude investigator bias. This might influence the NYHA stage, since investigator-dependent assessment tends to report better results. Therefore, comparison of functional outcome using NYHA classification is limited by the method of assessment and structured heart failure questionnaires are more reliable. However, as previously published we confirmed that functional outcome is improved by mitral TEER in both aFMR and vFMR. Furthermore, we found no significant difference in NYHA classification stage III/IV between vFMR and aFMR at 12-months follow-up. Also, dyspnea is multifactorial and the underlying cause of aFMR and vFMR might influence functional outcome.

One-year mortality was significantly lower in patients with aFMR compared to vFMR. 1-year survival accounts for 92.5% in the aFMR group and 81.8% in the vFMR group. Other studies reported no significant difference in mortality between aFMR und vFMR following mitral TEER.⁴ However, both a prospective international multi-ethnic cohort study and a metanalysis of current studies showed, that HFpEF has a better long-term survival than HFrEF, which strengthens our findings for aFMR and vFMR patients.^{13, 14}

Definition of aFMR is still a matter of debate and inclusion criteria vary among the studies.^{2, 15} Previous trials included only patients with AF in the aFMR group,^{5, 7, 12} while other also included HFpEF patients without AF.^{4, 11} The absence of diagnosed AF per se did not exclude aFMR, since LA dilatation could also be found in patients with sinus rhythm or in the presence of asymptomatic, not yet diagnosed paroxysmal AF. In the MITRA-PRO registry the aFMR group showed a history of AF in 83%, while AF was only assessed by the medical history and baseline or follow-up ECG. Therefore, one might speculate an underestimated rate of AF in the aFMR cohort. However, no adequate AF screening was performed in the MITRA-PRO registry. Therefore, AF was not used as an inclusion criterion for the aFMR group, which is in line with current reviews.^{2, 15}

Residual MR assessed by the MitraScore was an independent predictor of 1-year mortality in all FMR (n = 846; aFMR + vFMR) and vFMR (n = 722) patients. In aFMR, a MitraScore of ≥4 was associated with a higher mortality rate in univariate analysis. However, it was not an independent predictor of survival in multivariate analysis, likely due to the low event rate in this cohort. These findings highlight the critical role of LV function in overall survival for heart failure patients. Consequently, in vFMR patients, significant residual MR substantially impacts the 1-year mortality rate, whereas aFMR patients exhibit better tolerance to residual MR following mitral TEER. One might speculate that long-term follow-up (3 and 5 years) of the MITRA-PRO registry, which is not yet available, may reveal a MitraScore of ≥4 as an independent predictor of mortality following mitral TEER. However, at the 1-year follow-up, the MITRA-PRO registry only indicated a trend toward a higher mortality rate for aFMR patients with relevant residual MR. However, minimizing MR through mitral TEER to the lowest possible level is crucial for survival, especially in vFMR patients. In contrast, aFMR patients show better tolerance to significant residual MR 1 year after the initial procedure.

5 STUDY LIMITATIONS

Assessment of residual MR was done by MITRA-PRO investigators and investigator bias cannot be excluded. Furthermore, in this real-world registry guideline directed medical treatment (GDMT) was not controlled at 1-year follow-up. Therefore, changes in GDMT might have influenced the clinical outcome. Of note, the MITRA-PRO registry was designed before publication of COAPT and MITRA-FR study results and right ventricular function and LV dimensions were not considered for patient selection. MITRA-PRO included patients from 2016 to 2020, a period during which the novel MitraClip G4 was not yet available. This absence could potentially influence the incidence of residual MR. Furthermore, long-term follow-up (3 and 5 years) is yet not available and will provide further evidence of the impact of residual MR following TEER on long-term survival especially in patients with aFMR, where the event rate was low at 1-year follow-up. Logistic EuroScore, which is known to be an overpredictive risk model, was used in MITRA-PRO for risk assessment. EuroScore II or STS score would have been more accurate, but were not part of the study protocol designed in 2014. Mitral annular dimensions were not assessed in the MITRA-PRO registry and were not part of the inclusion criteria for aFMR, while LA dimensions were used to allocate patients to the aFMR group. AF might have influenced mortality of patients with aFMR and vFMR and was not assessed during the clinical follow-up. However, monitoring AF in clinical trials remains challenging, since ECG or even holter monitoring during follow-up might underestimate the burden of AF.

6 CONCLUSIONS

This MITRA-PRO registry analysis showed better 1-year survival in aFMR compared to vFMR regardless of relevant residual MR, while 1-year rehospitalization rates were comparable for both MR entities. Furthermore, aFMR patients had similar procedural outcomes regarding residual MR following TEER compared to vFMR patients, while the amount of implanted TEER devices was increased in vFMR. Residual MR is an independent predictor of 1-year mortality in vFMR patients, whereas in aFMR patients, a MitraScore of ≥4 is associated with higher mortality but is not an independent predictor in multivariate analysis. Therefore, minimizing MR through mitral TEER is crucial for survival in vFMR patients, while aFMR patients tolerate significant residual MR better 1 year after the procedure.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.