Long-term echocardiographic data, mechanisms of failure, and reintervention outcomes of the Epic valve in mitral position — a large observational cohort

,


CENTRAL MESSAGE
The Epic Mitral valve has stable hemodynamics over time and is associated with low incidence of failure and reintervention that is mostly due to endocarditis or leaflet tear without calcification.
PERSPECTIVE Large comparative long-term studies with other bioprostheses implanted in mitral position in terms of hemodynamics, rates and mechanism of failures, and reintervention outcomes are warranted.Reintervention type on failed Epic mitral valves seems to have no influence on early clinical outcomes, but larger multicenter studies of redo-MVR versus VinV in Epic Mitral valves are needed.
2][3] In recent decades, an important shift toward the use of biological over mechanical prostheses for MVR was observed. 4However, the main concern regarding bioprostheses in the prosthesis-selection process is their durability in the long term, especially in young patients, 5 but large echocardiographic reports on bioprostheses implanted in the mitral position are missing.The Epic stented porcine prosthesis (Abbott Structural Heart) is one of the most commonly used bioprostheses for MVR.It was commercialized in 2007 and is identical to its predecessor (Biocor) with the exception of having the Linx leaflet anticalcification treatment (ie, ethanol-based calcium mitigation therapy). 6Previous studies have assessed its early clinical performance in terms of survival and freedom from reintervention in large cohorts of patients who received the Epic in mitral and/or aortic position, [7][8][9] but long-term comprehensive echocardiographic follow-up (FU) reports of Epic Mitral valves are lacking to assess its performance in terms of valve hemodynamics.Moreover, although 2 small studies have assessed the Epic durability in the mid-term, 10,11 the independent determinants of reintervention or structural valve deterioration (SVD) remain unidentified.Also, rates of each mechanism of failures (SVD, non-SVD, endocarditis, thrombosis) differ or are missing in these studies.3][14] However, postreintervention outcomes exclusively on failed Epic prostheses according to reintervention approaches (redo-MVR or TMVR-VinV) are currently unknown.There is thus a need for long-term clinical and echocardiographic data in a large cohort of Epic Mitral valves to address these gaps and discrepancies in the literature. 15e aimed to (1) evaluate the long-term valve hemodynamics and durability of the Epic Mitral valve; (2) assess the rates, mechanisms, and independent predictors of Epic prostheses failures; and (3) to compare the short-and mid-term outcomes according to reintervention type.

METHODS Ethical Statement
The study was approved by the local ethics committee (Institut universitaire de cardiologie et de pneumologie de Qu ebec-Universit e Laval, 2022-01-20, #860695) and, due to its retrospective nature, patient consent was waived.

Study Population
This study includes all consecutive patients who underwent an MVR and received the Epic stented porcine xenograft using the Linx leaflet anticalcification treatment from November 2007 to May 2021 at the Institut universitaire de cardiologie et de pneumologie de Qu ebec-Universit e Laval.Patients included in this registry were 18 years or older with either an isolated MVR or in combination with coronary artery bypass grafting (CABG), atrial septal defect or foramen ovale closures, maze procedures, aortoplasty, and/or other valvular procedures.The exclusions criteria were any adult congenital heart diseases interventions or the following concomitant procedures: myectomy, aneurysm resection, ventricular septal defect closure, vascular bypass intervention, and pulmonary aortoplasty.

Procedures and Follow-up
Patients underwent standard MVR procedures with median sternotomy and cardiopulmonary bypass.The decision regarding bioprosthesis valve type selection was individualized and left to the discretion of the surgeon.Details regarding surgical procedures and postoperative medical treatment appear in Appendix E1.All pre-, intra-, and postoperative clinical, laboratory, or echocardiographic data were retrieved from our institutional surgical database that collects patients' data consecutively.
Routine FU included clinical evaluation and transthoracic and/or transesophageal echocardiography at discharge, 6 to 12 months, and then yearly up to 5 years of FU.Patients completed their echo FU in our center or referring hospital based on geographic distribution, and efforts were deployed to retrieve external echocardiographic data using a standardized approach.The main valve hemodynamics parameters of interest during FU were (1) peak and mean gradients (PG; MG); (2) effective orifice area (EOA), derived from the continuity equation; and (3) at least moderate mitral regurgitation (MR), defined as either central or paravalvular leak (PVL).Clinical outcomes included 30-day mortality and early postoperative adverse events (eg, stroke, deep sternal would infection, reoperation for bleeding, and so on), mitral valve reintervention, and all-cause mortality.Survival data and reintervention occurrence were derived from an indepth interrogation of medico-administrative governmental data.

Mechanism of Failures
The causes of bioprostheses failures were established through a standardized methodology by both an experienced echocardiographist and surgeon according to known and standardized definitions of bioprosthesis valve dysfunction and bioprosthesis valve failure in the mitral position (details in Appendix E1). 16

Reintervention
The choice of reintervention approach was left to the decision of the heart team according to the mechanism of failure and patients characteristics.The reintervention was accomplished either by reoperation for MVR (redo-MVR), TMVR-VinV procedure, or PVL closure intervention (surgical or transcatheter; if that was the only observed mechanism of failure).One patient underwent a successful thrombectomy procedure without any need of further valve intervention.Postreintervention outcomes were either derived from our surgical or cathlab research team institution's databases.

Statistical Analysis
Results are expressed as standard descriptive statistics: absolute and relative frequencies for categorical variables and mean AE standard deviation or median and interquartile range for continuous data.Data were compared between groups (ie, mechanism of failures and reintervention approach) using c 2 or the Fisher exact tests for categorical variables and the Student t-test or analysis of variance and their respective nonparametric tests (Mann-Whitney or Kruskal-Wallis) for continuous variables.Bonferroni or C de Dunnett post-hoc tests were conducted for multiple comparison.Evolution of MV hemodynamics (ie, PG, MG, EOA) during FU were compared with paired t-test, or Wilcoxon signed-rank test, and McNemar-Bowker's analyses, as appropriate.For the latter, patients who underwent reintervention during the 5-years echocardiographic FU were censored at the last FU available before reintervention.Kaplan-Meier analyses were conducted to derive event-free survival rates from all-cause MV reintervention and reinterventions due to each mechanism of failure (ie, SVD, non-SVD, and endocarditis and thrombosis combined) and to compare all-cause mortality between reintervention approach (ie, redo-MVR or VinV).Cumulative incidence function was used to assessed the potential competitive risk of death and no effect was observed on incidences of all-cause MV reintervention.Univariate and multivariate Cox analyses were conducted to determine the independent predictors of (1) mortality, (2) all-cause reintervention, and (3) SVD-related reintervention.Variables with a P .10 in univariate analyses or with known associations with the outcomes of interest were used for multivariate analyses.This was achieved with a careful inclusion of variables to avoid collinearity and with a preference for continuous variables over dichotomic (eg, continuous left ventricular ejection fraction [LVEF] rather than LVEF <50%).Data were analyzed using SPSS software (version 28, IBM Corp).

Study Population and Baseline Characteristics
Of the 2960 patients who underwent a MVR at our institution between 2007 and 2021 and who respected our inclusion criteria, 1397 (47%) patients received an Epic prosthesis.Of those, between 13% and 39% had an echocardiographic FU at the predefined time points (ie, 1-3 months, 6 months, yearly; Figure 1), which generated a total of 1021 (73%) patients with at least 1 echocardiographic FU.
Twenty-one percent of patients had an isolated MVR and the main concomitant procedure was CABG (46%) with concomitant aortic or tricuspid valve intervention accomplished in 28% and 18% of cases, respectively (Table 1).The main Epic prosthesis sizes implanted were 29 and 31 mm in 28% and 31% of cases, respectively (Table 1).

Early Postoperative Data
The 30-days mortality is 7% (n ¼ 94) and the median length of stay is 9 days for all cohort.However, the 30day in-hospital mortality was 3.5% (n ¼ 10) for isolated MVR (n ¼ 288) with a median hospital stay of 8 days (Table E1).At discharge, the mean PG and MG gradients were 16 AE 5 and 6 AE4 mm Hg, respectively, and 8 patients (0.6%) had moderate or greater residual MR, which were all PVL (Table E1).Three of them had asymptomatic moderate-to-severe, and 5 had moderate PVL.

Epic Mitral Valve Hemodynamics During Follow-up
PG, MG, and EOA of the Epic prostheses were stable over time (Figure 2, A and B).Mean values of these valve hemodynamics parameters at each FU with paired analysis are provided in Table E2.Although some significant paired differences were observed, the differences were always in the right direction (eg, lower gradients [or greater for EOA] in the later FUs compared with the earlier FUs, all details appear in Table E2).In brief, these significant paired differences for MV hemodynamics were (1) for PG: only between 1 and 3 months and 3 years' FUs (P ¼ .04);(2) for MG: between 1 and 3 months' FUs and all time points (all P .04),except with 4 years' FUs, and between 6 months and 3 years' FUs (P ¼ .02);and (3) for EOA: between 1 and 3 months and 3 years' FUs (P ¼ .04)and between 2 and 5 years' FUs (P ¼ .03).Incidences of MG !10 mm Hg through the 5 years' FU were between 4% and 7% (Table E3), whereas residual/ recurrent moderate or greater MR (ie, PVL or central) were between 0.5% and 4% for all time points (Table E3).The Journal of Thoracic and Cardiovascular Surgery c December 2023 Adult: Mitral Valve Bernard et al

ADULT Reintervention and Mechanisms of Failures of Failed Epic Mitral Prostheses
A total of 70 (5%) patients underwent a MV reintervention at a median time-to-reintervention of 3.0 (0.7-5.4) years by either redo-MVR, VinV, or PVL closure in 54%, 27%, and 17% (surgical: n ¼ 3 and transcatheter: n ¼ 9), respectively (Figure 1).Of the 70 cases, 53% were female with a mean age at reintervention of 70 AE 10 years, and the main Epic Mitral valve sizes that failed were 29 and 31 mm in 26% and 31% of cases, respectively (Table 2).All patients with nonendocarditis or thrombose-related failed Epic (n ¼ 42) presented dyspnea secondary to severe MR.
The interval between onset of symptoms and reintervention was acute (<1 month) or subacute (1-3 months) in 31 patients and chronic (>3 months) in 11 patients.
The rates of each mechanism of failures are presented in Figure 2, C. SVD was present in 27 cases, which represent a 1.9% incidence in the total population at a median time-toreintervention of 6.0 (4.1-9.0)years.All SVD cases were due to leaflet tear, with no calcification found during the intraoperative assessment.The second most frequent mechanism of failure was endocarditis (n ¼ 24), with 1.7% incidence at a median time-to-reintervention 0.8 (0.5-1.8) years.SVD cases had initially a 31-mm prosthesis (46%) and were mainly retreated by VinV (69%; Table E4).Differences between each mechanism of failure are presented in Table E4 and can be all attributed to either the inherent mechanism (eg, endocarditis for greater EuroSCORE II) or the reintervention approach (eg, for longer intensive care unit stay).
Freedom from all-cause MV reintervention and related to each mechanism of failures are presented in Table 3.Of note, freedom from all-cause MV reintervention at 5 and 10 years were 95% and 89%, respectively (Table 3; Figure 2, D), and 99% and 92% for SVD-related reintervention, respectively (Table 3).
During a median FU of 4.5 (2.7-6.9)years, there were 417 deaths (10-year survival: 51 AE 3%).After comprehensive multivariate analysis, all-cause MV reintervention on failed Epic as a time-dependent variable was associated with a $2.1-fold increased risk of long-term all-cause mortality (P ¼ .02;Table E5).Univariate associations of MV reintervention due to each mechanism of failure (as time-dependent variables) with late mortality are provided in Table E5.

Determinants of Epic Valve Reintervention
Univariate determinants of all-cause MV reintervention were age, renal clearance, AF, coronary artery disease, active endocarditis, concomitant procedures (above all: concomitant CABG), and etiology of MV disease (eg, above all: endocarditis and mitral stenosis) at the initial surgery, and moderate or greater PVL at discharge (Table 4).Other variables also presented strong tendencies for greater risk of MV reintervention: baseline obesity, concomitant aortic or tricuspid valve intervention, and de novo AF at 30 days' postoperatively (Table 4).After multivariate analysis, only age, baseline AF, initial MV etiology, and moderate or greater PVL at discharge (all P .04;Table 4) were predictive.Moreover, multivariate models were conducted with each MV etiology subtype and revealed that only the initial mitral stenosis was an independent predictor of all-cause MV reintervention on the Epic (Table E6).
Regarding SVD-related reintervention, the univariate predictors were only age, EuroSCORE II, renal clearance and AF at the initial surgery, and moderate or greater aortic regurgitation (AR) at discharge (Table 4).Although limited in statistical power, in multivariate analyses adjusted for EuroSCORE II (which includes age and renal disease), baseline AF and moderate or greater AR at discharge remained significantly associated (all P .02;Table 4).

Comparison of Outcomes According to Reintervention Approaches
Of the 38 (54%) patients who underwent redo-MVR, 55% of them were female compared with 37% in VinV (P ¼ .19).The 2 reintervention approaches significantly differed in terms of age at reintervention, time-toreintervention, and mechanisms of prosthesis failure, as expected (Table 2).The main replacement prosthesis that was reimplanted by redo-MVR and VinV procedures was the Epic in 63% and the SAPIEN 3 in 95% of cases, respectively (Table 2).Regarding adverse events at 30 days and echocardiographic data at discharge, no significant differences were observed except for more low cardiac output state and longer total length/intensive care unit stays in redo-MVR (all P .01),as expected (Table 2).During a median FU of 2.5 (1.2-4.3) years after reintervention, there were 19 deaths (27% of all failures), which 14 of them were in redo-MVR or VinV cases and 5 in the PVL closure cases (1 surgical and 4 transcatheter [2 failed attempt]).Comparison of survival up to 3 years postreintervention demonstrated no significant difference between reintervention type (univariate and age-adjusted P !.34; Figure 3).

DISCUSSION
In this large and comprehensive cohort of patients, we aimed to describe the long-term durability of the Epic prosthesis implanted in the mitral position and the reintervention outcomes of failed Epic Mitral valves by the means of assessing (1) the variations of valve hemodynamic parameters over 5 years of echocardiographic FU, (2) the rates and the freedom of reintervention for each mechanism of failure (SVD, non-SVD, endocarditis, or thrombosis), (3) the independent predictors of failed Epic Mitral valves, and (4) the comparison of reintervention approach (ie, redo-MVR and VinV).The main findings are (1) the Epic Mitral valve has stable gradients and EOA over time; (2) the rates of SVD and all-cause reintervention over 10 years of follow-up are 1.9% and 5%, respectively, which leads to a freedom from all-cause and SVD-related MV reintervention at 10-years of 92% and 89%; (3) the main predictors of valve failure are age, baseline AF, initial MV etiology, and moderate or greater PVL at discharge; and (4) VinV and redo-MVR approaches on failed Epic Mitral valves provided similar early clinical outcomes and 3-year mortality (Figure 4).

Current State of Evidence on Durability of the Epic Valve
The Epic is a third-generation prosthesis composed of 3 natural porcine leaflets with the Linx leaflet anticalcification treatment.This improvement was applied to optimize the durability of the prosthesis in the long term, 17 but large reports with longer clinical and echocardiographic FUs are needed to establish reliable rates of prosthesis deterioration and assess its durability in the long-term for patients outcomes. 158][9] In these previous studies, the number of patients who received the Epic in the mitral position varies between 14% and 23%.Because of this limitation and because the rates of each mechanism of failures differ or are missing in these reports, it is difficult to establish some reference values of mechanism of failures to guide bioprosthesis selection in patients who undergo MVR.Another limitation of these studies is the small sample size, which limits the authors' ability to accomplish some of our study objectives: (1) compare characteristics between mechanism of failures, (2) determine independent predictors of reintervention, (3) establish the associations of each mechanism of failures with death, and (4) compare the outcomes of reintervention approaches.Regarding freedom from SVD-related reintervention, Anselmi and colleagues 11 recently stratified the freedom from SVD-related reintervention on the Epic according to age and found no intergroup difference.Yet, they identified only 9 (1.9%)SVD cases in a younger population, which represent an important statistical power limitation to identify significant associations with potential predictors of SVD.Lastly, a recent study compared the modes of prosthesis failures between the Carpentier-Edwards PERIMOUNT pericardial and the Epic Mitral prostheses and found that the Carpentier-Edwards PERIMOUNT deteriorated more of leaflet calcification versus leaflet tearing in Epic, which is consistent with our results. 18

Importance of Hemodynamic Performance Assessment
No studies to date have assessed the hemodynamics of the Epic Mitral valve at mid-or long-term, even if these parameters are paramount to assess the bioprosthesis performance  bioprosthesis-selection process.0][21][22][23][24] However, there is a need for more studies assessing the VinV suitability and outcomes compared with redo-MVR in each type of surgical bioprostheses, because the outcomes are expected to be different according to valve type and conceptual design.In the present study, we have addressed this lack of knowledge on the Epic prosthesis.We observed no differences in terms of (1) early adverse events, except for the predictable ones (eg, length of stay); (2) hemodynamics at discharge; and (3) midterm mortality.However, these patients should be addressed by an individualized approach in a multidisciplinary heart team because several factors need to be considered such as the mechanism of failure (eg, redo-MVR for endocarditis), the failed prosthesis type and size, patients' age, symptomatic status, comorbidities, [25][26][27][28] and more, as redo-MVR can provide excellent long-term results in selected cases. 29,30Large comparative studies with long-term echocardiographic FU are needed before expanding indications of mitral VinV. 31,32udy Strengths and Limitations This is a retrospective study and is thus subject to its inherent limitations, but it represents the largest singlecenter study to date on Epic prostheses implanted in the mitral position.Echocardiographic FU was not available in all patients at the predefined time points and may be susceptible to interobserver variations, but 70% of patients included had at least 1 echocardiographic FU up to 5 years, and paired analyses were used to assess the changes in valve hemodynamic parameters between FU.This kind of comprehensive echocardiographic report is not yet available for any other bioprosthesis implanted in the mitral position, even if this end point is the cornerstone of the durability assessment of heart valve prostheses.This study has no subgroup comparisons for hemodynamics, but it provides accurate rates of all mechanisms of Epic failures.Although limited in terms of redo-MVR and VinV cases, this study is the first to describe the reintervention outcomes up to 3 years in a dedicated mitral porcine prosthesis, as other studies that compare redo-MVR and VinV use a grouping of various types of surgical valves.This current lack of knowledge limits the assessment of the suitability of each surgical mitral prosthesis type for future VinV.The low incidence of SVD-related reinterventions limits our analyses to ascertain independent predictors due to low statistical power, but future studies should investigate potential predictors of subclinical SVD and its associations with outcomes.Nevertheless, this represents the first largest study that addresses both the long-term Epic mitral valve hemodynamics and independent predictors of failures.

CONCLUSIONS
The Epic mitral valve has stable hemodynamics through 5 years and is associated with an incidence of 1.9% of SVD and 5% of all-cause reintervention through 10 years of follow-up.The primary mechanism of SVD in Epic Mitral valves is leaflet tear without calcification.Independent determinants of reintervention are more related to patient's characteristics than the prosthesis itself aside from initial MV etiology and residual paravalvular leak at discharge.Reintervention type on the Epic Mitral valve has no influence on short-and mid-term outcomes.Large comparative studies with other bioprostheses implanted in the mitral position are warranted.
Conflict of Interest Statement P.P. received funding from Edwards Lifesciences, Medtronic, and Pi-Cardia for echocardiography core laboratory analyses, with no personal compensation.M-A.C. received funding from Edwards Lifesciences for computed tomography core laboratory analyses with no direct personal compensation and a research grant from Medtronic.S.M. is consultant for Abbott, Edwards Lifesciences, and Medtronic.All other authors reported no conflicts of interest.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict

Clinical Implications
The Epic Mitral valve has stable hemodynamics through 5-years and is associated with an incidence of SVD and all-cause reintervention through 10-years of FU of 1.9 and 5%, respectively.Reintervention type on failed Epic Mitral valves has no influence on early clinical outcomes and 3-year mortality.The surgeon performed a preservation of the subvalular apparatus during mitral valve replacement (MVR) if possible, according to the presentation of the mitral disease at the time of the surgery.Operative risk was assessed using the European System for Cardiac Operative Risk Evaluation II criteria for all patients initially treated with MVR and at the time of reintervention for comparison between redo-MVR and valve-in-valve.Postoperative anticoagulation therapy was prescribed using warfarin for 3 months.The choice of antithrombotic regimens was subject to change at the discretion of the surgeon or referring cardiologist during follow-up.
The echocardiographic parameters provided in this study were measured according to current American Society of Echocardiography guidelines at the time of the procedure.E1,E2

Mechanisms of Failure
Based on recent standardized criteria for assessment of bioprosthesis valve dysfunction and failure in the mitral position, E3 the mechanisms of failure were established as follows: structural valve deterioration (SVD), non-SVD, valve endocarditis, and valve thrombosis.The following diagnosis criteria were used to establish SVD using transthoracic and/or transesophageal echocardiography: visual confirmation of leaflet tearing or disruption, prolapse/ flail, leaflet fibrosis or calcification, or severe regurgitation (effective regurgitant orifice area >40 mm 2 , regurgitant volume>60 mL).Regarding non-SVD, the criteria were evidence of significant paravalvular leak, severe prosthesispatient mismatch or leaflet entrapment by pannus or suture, and inappropriate positioning or embolization of the valve.Finally, thrombosis occurrence was detected only by echocardiography (not confirmed by computerized tomography) and endocarditis was defined by evidence of abscess, pus or vegetation confirmed in most cases by histological or microbiological studies.

FIGURE 1 .
FIGURE 1. Study flow diagram.This figure shows the number of patients who had echocardiographic FU at the predefined time points and rates of each reintervention approaches with their respective echocardiographic FU postreintervention.MVR, Mitral valve replacement; IUCPQ-UL, Institut universitaire de cardiologie et de pnemologie de Qu ebec-Universit e Laval; FU, follow-up; MV, mitral valve; TMVR, transcatheter mitral valve replacement; V-in-V, valve-in-valve.

FIGURE 2 .
FIGURE 2. Epic Mitral valve hemodynamics, mechanisms of failures and reintervention throughout FU.Evolution of peak and mean gradient (A) and EOA (B) from discharge to 5 years' FU. (C) Rates of each mechanism of failure of the Epic Mitral valve in the total population.(D) Freedom from allcause MV reintervention throughout 12 years of clinical FU with 95% confidence intervals as blue shading bars.*One patient had both SVD and thrombosis; thus, the sum of the rates can be>100%.This patient was excluded for further analyses.The percentages depicted in (D) show the 6 and 12 years' freedom from reintervention.EOA, Effective orifice area; SVD, structural valve deterioration; MV, mitral valve; PVL, paravalvular leak.
Consecutive Mitral Valve Replacement with the Epic TM Between November 2007 and May 2021

TABLE 3 .
Freedom from reintervention on Epic Mitral prosthesis The event-free survival obtained by Kaplan-Meier analyses at a given time is depicted as value AE standard error (SE).SVD, structural valve deterioration.MG >10 mm Hg and moderate or greater PVL or central MR.Although these results demonstrate the competence of the Epic to manage MV inflow, there may be certain cases in which the leaflet may be more susceptible to leaflet tearing.As a matter of fact, one of the strong predictors of Reintervention Approach for Failed Epic Valves Suitability of surgical valves for future VinV is an important consideration to take into account in the

TABLE 4 .
Determinants of all-cause and SVD-related reinterventions E1. Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, et al.Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging.J Am Soc Echocardiogr.2015;28: 1-39.E2.Zoghbi WA, Asch FM, Bruce C, Gillam LD, Grayburn PA, Hahn RT, et al.Guidelines for the evaluation of valvular regurgitation after percutaneous valve repair or replacement: a report from the American Society of Echocardiography developed in collaboration with the Society for Cardiovascular Angiography and Interventions, Japanese Society of Echocardiography, and Society for Cardiovascular Magnetic Resonance.J Am Soc Echocardiogr.2019;32:431-75.E3.Pibarot P, Herrmann HC, Wu C, Hahn RT, Otto CM, Abbas AE, et al.Standardized definitions for bioprosthetic valve dysfunction following aortic or mitral valve replacement: JACC State-of-the-Art review.J Am Coll Cardiol.2022;80: 545-61.

TABLE E2
Values in bold indicate statistical significance.MV, Mitral valve; FU, follow-up; EOA, effective orifice area.
The Journal of Thoracic and Cardiovascular Surgery c Volume 166,Number 61655.e4