Long-term outcomes after bioprosthetic tricuspid valve replacement: a multicenter study

Abstract

OBJECTIVES

Long-term evidence about bioprosthetic tricuspid valve replacement is scarce. This study aims to investigate the long-term clinical outcomes of patients who underwent tricuspid valve replacement with bioprostheses.

METHODS

This multicentre retrospective study included patients from 10 high-volume centres in 7 different countries, who underwent tricuspid valve replacement with bioprostheses. Echocardiographic and clinical data were reviewed. Long-term outcomes were investigated using Kaplan–Meier estimates, Cox regression, and competing risk analysis.

RESULTS

Of 675 patients, isolated tricuspid valve replacement was performed in 358 patients (53%), while 317 (47%) underwent concomitant procedures. Between these 2 groups, patients who underwent combined procedures reported a significantly higher incidence of infection, atrioventricular block, multi-organ failure, longer intensive care unit and hospital stay and higher 30-day mortality over patients who underwent isolated procedure. The overall 30-day mortality occurred in 70 patients (10.4%) [46 (14.6%) combined vs 24 (6.74%) isolated, P = 0.001]. During the follow-up, there was a continuous rate of attrition due to death, with cumulative incidences of death at 5, 10 and 15 years being 27.2%, 46.2% and 60.6%, respectively. In contrast, the risk of reintervention starts to significantly increase after 10 years of follow-up, with cumulative incidences of reintervention being 6.1%, 10.8% and 23.3%, respectively. Freedom from tricuspid valve reintervention, pacemaker implantation, tricuspid valve endocarditis and major thromboembolic events at 15 years were 56.5%, 77.3%, 84.0% and 86.4%, respectively.

CONCLUSIONS

Tricuspid valve replacement with bioprostheses is an effective treatment for valvular disease, despite being associated with relatively high early and long-term mortality. However, the risk of structural valve degeneration rises significantly after 10 years.

Open in new tabDownload slide

INTRODUCTION

Tricuspid valve replacement (TVR) with bioprostheses is generally performed as a concomitant procedure during left heart valve surgery, while isolated TVR is performed less frequently [1, 2]. TVR is usually reserved for patients in whom a tricuspid valve repair is not feasible or when repair attempts have failed. There is still discussion about the best valve substitute or the right timing to perform surgery. Although robust data are available about short- and long-term results following tricuspid valve repair, evidence about the durability, freedom from reoperation and clinical outcomes of patients with a bioprosthesis in a tricuspid position is scarce. Hence, the purpose of this multicentre retrospective study is to analyse the long-term outcomes of patients who underwent bioprosthetic TVR.

METHODS

Ethical statement

Every participating centre approved the study design, consent process and review and analysis of the data (IRB approval CER-BDX 2025–49). We also guarantee the respect of anonymity, professional secrecy and the use of the collected data and the statistical analysis solely for the scientific purposes granted in accordance with the law in force.

Study population

A total of 675 consecutive patients from 10 centres who underwent bioprosthetic TVR from 2003 to 2023, were enrolled in this study. The participating centre sites along with the numbers of enrolled patients are depicted in Fig. 1.

Patients older than 18 years who underwent bioprosthetic TVR (also for failed repair), with or without additional concomitant cardiac surgical procedures were included in the analysis.

Pre- and postoperative echocardiographic evaluations were performed by board-certified cardiologists according to international guidelines in force during the study period [3].

The indication for surgery was tricuspid valve regurgitation (TR) (degenerative and functional) and endocarditis. All types of bioprostheses used by the participating centres were taken into consideration. The indication for surgery, the adopted repair or replacement strategy, the choice of implanted bioprosthesis and other technical details depend on the preferences of individual centers, institutional practices and the experience of the surgeons. None of the included patients underwent minimally invasive procedures.

Data were collected from the institutional electronic health records. The follow-up data were collected during the follow-up period directly at the participating institutions or by telephonic survey for missing patients. The follow-up was closed in December 2023. Morbidity events were defined according to the ‘Guidelines for Reporting Mortality and Morbidity After Cardiac Valve Interventions’ [4, 5] written by the Society of Thoracic Surgeons (STS), the Councils of the American Associations of Thoracic Surgery (AATS), and the European Association of Cardio-Thoracic Surgery (EACTS).

The study primary outcomes were as follows: the long-term survival and the freedom from TV reintervention. The secondary end-points were as follows: the freedom from endocarditis, new pacemaker implantation and right ventricular (RV) failure needing hospitalization.

Statistical analysis

Continuous variables were evaluated for normality using the Shapiro–Wilk test and reported as mean (standard deviation) or median (interquartile range), as appropriate, while categorical variables were reported as frequencies and percentages.

Kaplan–Meier curves and estimates were generated for the following time-to-event outcomes: overall survival, freedom from TV reintervention, freedom from pacemaker implantation, freedom from TV endocarditis, freedom from hospitalization for RV failure, and freedom from major adverse events. The log-rank test was used when comparisons between groups were made. In addition, we performed a competing risk analysis to estimate the probabilities of the following three outcomes ‘reintervention’, ‘death’ and ‘alive without reintervention’. We used the ‘cmprsk’ function in R, which estimates the cumulative incidence function of the event of interest (i.e. ‘reintervention’) while taking into account competing events (i.e. ‘death’). The probabilities of ‘reintervention”’and ‘death’ were visualized using cumulative incidence curves, while the probability of being ‘alive without reintervention’ was visualized using a Kaplan–Meier curve.

Completeness of follow-up was expressed by means of the follow-up that measures the ratio between the actual follow-up period (in patient-years) and the potential follow-up period (in patient-years). It helps in understanding how much of the intended follow-up was achieved. The potential follow-up was defined as the amount of patient-years from the index TVR procedure to death or locking of the dataset on 31 December 2023.

In order to identify predictors of mortality and TV reintervention, Cox and Fine & Gray regression models were constructed, respectively, producing hazard ratios (HRs) and subdistribution hazard ratios (SHRs), respectively, with 95% confidence intervals.

First, univariable models were tested for each of the predictor variables. Then, multivariable models were constructed; starting from the null model, variables were introduced in a stepwise fashion based on the Akaike information criterion (AIC). All models were created using the R package ‘gtsummary’ and were checked for overfitting. Before creating the regression models, missing data were imputed using a random forest model via the R package ‘missForest’. All analyses were completed with R Statistical Software (version 4.1.1, Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Study population

The median age at the procedure was 66 years, and 46% were male. About half of patients (47.6%) was functioning in New York Heart Association (NYHA) class III, about one-third (34.2%) in NYHA class II. The majority of patients (81.2%) had a TR grade of ≥3. Isolated TVR was performed in 358 patients (53%), while 317 (47%) underwent concomitant procedures. Beating heart TVR was performed in 176 (33.4%) patients of the entire cohort, 110 (62.5%) of them as isolated TVR, while the remaining 37.5% as combined procedures. Other demographics and preoperative characteristics are summarized in Table 1.

In-hospital outcomes

Early postoperative outcomes are summarized in Supplementary Material, Table S1. The 30-day mortality for all patients was 10.4%, and it was higher following combined procedures compared to isolated TVR (14.6% vs 6.7%, P = 0.001). The median intensive care unit and hospital lengths of stay were 5 days (3–11) and 17 days (11–27), respectively, and were significantly shorter for isolated TVR (both P < 0.001). The most common postoperative complications were infection (38.4%), including pneumonia, urinary tract infections, sternal wound infections, atrioventricular block requiring pacemaker implantation(25.9%), new atrial fibrillation(19.0%), acute kidney injury requiring dialysis (17.0%) and multiorgan dysfunction syndrome (14.9%); each of these complications was more common after combined procedures (all P = 0.030). At discharge, the mean TV gradient was 4 mmHg and paravalvular leakage was observed in 6.3% of patients. Finally, 79.4% of patients were completely free from TR at discharge, while 19.4% had only mild residual TR.

Long-term mortality and TV reintervention

Data on mortality were available in all 675 patients. A total of 217 patients were alive and being followed up for at least 5 years (corresponding to 1695 actual patient-years out of 2386 potential patient-years, resulting in a follow-up index of 71.0%), 58 patients for at least 10 years (corresponding to 2307 actual patient-years out of 3561 potential patient-years, resulting in a follow-up index of 64.8%) and 5 patients for at least 15 years (corresponding to 2431 actual patient-years out of 3889 potential patient-years, resulting in a follow-up index of 62.5%). Kaplan–Meier estimates for overall survival are depicted in Fig. 2A. Overall survival estimates at 5, 10 and 15 years were 69.0%, 47.2% and 28.9%, respectively. Data on TV reintervention were available for 490 patients, with a median follow-up of 2.0 years (0.21–6.0 years).

During follow-up, 41 patients underwent any cardiac surgical reoperation and 31 of them a TV surgical reintervention. These were 19 Edward’s Magna Ease, 8 Medtronic Epic, 1 Medtronic Mosaic, 2 Medtronic Hancock and 1 not known. In these 31 patients, the mechanism of valve degeneration/dysfunction was moderate/severe valve regurgitation in 5 patients, moderate/severe valve stenosis in 11 patients, mixed regurgitation/stenosis in 3 patients, 1 case of severe prosthetic valve leak and 12 endocarditis. Kaplan–Meier estimates for TV reintervention are shown in Fig. 2B. Estimates for freedom from TV reintervention at 5, 10 and 15 years were 92.5%, 83.9% and 56.5%, respectively. The competing risk analysis of TV reintervention and death is presented in Fig. 3. There was a continuous rate of attrition due to death during follow-up, with cumulative incidences of death being 27.2% at 5 years, 46.2% at 10 years and 60.9% at 15 years. In contrast, the risk of TV reintervention remained low until 10 years and then started to increase, with cumulative incidences of TV reintervention being 6.1% at 5 years, 10.8% at 10 years and 23.3% at 15 years. At any point in time, patients were more likely to have died than to have received a TV reintervention. The probability of being alive without reintervention decreased from 66.6% at 5 years to 43.0% at 10 years, and only 16.1% at 15 years. Figure 4A and B provides stratified analyses of overall survival and freedom from TV reintervention according to the following groups: combined procedures and isolated TVR procedures. The log-rank test indicated no significant difference in terms of long-term overall survival rates (P = 0.095), whereas the rates of TV reintervention were higher in those who underwent isolated TV reintervention (P = 0.009).

Predictors of mortality

Regression modelling revealed older age (HR 1.02 per year), higher AST (HR 1.61 per log-transformed U/l), higher MELD-Xi (HR 1.01 per point increase) and urgent or emergent surgery to be independent risk factors for mortality (Supplementary Material, Table S2). Supplementary Material, Tables S3 and S4, present the Fine & Gray and Cox models, respectively, used to investigate the risk factors for TV reintervention. Interestingly, older age (HR 0.97 per year, SHR 0.97) and preoperative atrial fibrillation (HR 0.19, SHR 0.28) were identified to be protective factors with regard to risk of TV reintervention.

Additional outcomes

A total of 144 patients (22.1%) already had a pacemaker preoperatively. Among the remainder of the patients, 295 had data available on the occurrence of new pacemaker implantation postoperatively, with a median follow-up of 1.8 years (0.14–5.9 years). A total of 42 patients received a new pacemaker during follow-up with a cumulative incidence at 5, 10 and 15 years of 17%, 22.7% and 22.7%, respectively. Kaplan–Meier estimates are given in Fig. 2C. Estimates for freedom from new pacemaker implantation at 5, 10 and 15 years were 83.0%, 77.3% and 77.3%, respectively.

Data on TV endocarditis were available for 371 patients, with a median follow-up of 2.0 years (0.21–6.2 years). During the follow-up period, a total of 21 patients developed TV endocarditis (10 Magna Ease, 8 Epic, 1 Mosaic, 1 Hancock, 1 not known), 12 of them required TV reoperation (5 Magna Ease, 5 Epic, 1 Hancock, 1 Mosaic), while 9 patients were conservatively managed. Kaplan–Meier estimates are given in Fig. 2D. Estimates for freedom from TV endocarditis at 5, 10 and 15 years were 93.9%, 87.8% and 84.0%, respectively.

Data on hospitalization for RV failure were available for 368 patients, with a median follow-up of 2.0 years (0.20–6.2 years). A total of 37 patients were required to be hospitalized for RV failure during follow-up. Kaplan–Meier estimates are given in Fig. 2E. Estimates for freedom from hospitalization for RV failure at 5, 10 and 15 years were 88.1%, 80.0% and 74.0%, respectively.

Data on major adverse events were available for 369 patients, with a median follow-up of 2.0 years (0.19–6.2 years). A total of 28 patients developed major adverse events during follow-up. No case of valve thrombosis has been reported. Kaplan–Meier estimates are given in Fig. 2F. Estimates for freedom from major adverse events at 5, 10 and 15 years were 92.9%, 86.4% and 86.4%, respectively.

DISCUSSION

This multicentre study evaluated the clinical and haemodynamic outcomes of bioprostheses used for TVR over 2 decades in 10 Institutions. During follow-up, a gradual upward trend in the cumulative incidence of mortality was evident. Simultaneously, the risk of reoperation remained relatively low for the initial 10-year span, after which it demonstrated a notable increase. Moreover, the probability of being alive without any cardiac reintervention exhibited a discernible decline from 66.6% at the 5-year mark to 43% at 10 years, and ultimately, 16.1% at 15 years. These data serve as a benchmark for patients, cardiologists and surgeons to assist with decision-making about the use of bioprostheses for TVR, giving a comprehensive summary of the risks of mortality, reintervention, valve degeneration/dysfunction, endocarditis, and major events.

Early and long-term prognosis of patients undergoing TV surgery generally remains poor, with high morbidity and mortality [6–8]. Although TV repair should be preferred whenever feasible [9, 10], significant concerns about TVR persist, including optimal prosthesis selection, reoperation risks and future patient outcomes.

The cohort analysed in this study had an overall calculated EUROscore II of 6%, with 5.6% for isolated procedures and 6.2% for combined procedures. The results showed an overall early mortality of 10.4% (6% for isolated procedures and 14% for combined procedures). This finding is consistent with previous studies [9, 11, 12] and further emphasizes that patients undergoing tricuspid valve surgery are high-risk and highly fragile. Moreover, particularly for concomitant procedures, there may be risk factors that current risk scores fail to adequately capture.

The preference for tissue over mechanical valves for TVR is influenced by the lower incidence of thrombo-hemorrhagic events, the potentially greater long-term durability due to lower stress in the right-sided cardiac chamber, and the typically shorter long-term survival of patients with TV disease due to pulmonary hypertension, comorbidities and frequent RV dysfunction [1–14]. However, a recently published meta-analysis demonstrates how mechanical prostheses are associated with better long-term survival and lower risk of reoperations when compared with bioprostheses [15]. This meta-analysis shows how the rate of reoperation rapidly increases 7 years after TVR with tissue valvular substitute [1, 16–18], dispelling the belief that bioprostheses in the tricuspid position degenerate more slowly than those in the aortic or mitral position [19–21]. Previous experience with bioprostheses for TVR reported a freedom from SVD ranged from 97% to 68% at 10 years [19, 22–24]. Consistent with these results, our study reports freedom from tricuspid valve reoperation of 83.4% at 10 years. Nevertheless, the rate of any cardiac reoperation increases rapidly beyond 10 years after the initial procedure, and less than half of the patients (43%) are alive without the need for a cardiac reoperation at 10 years.

A sub-analysis comparing isolated TVR and combined TVR shows a significantly higher early mortality (14.6% vs 6.7%, P = 0.001) accompanied by a significantly higher incidence of infections (43.2% vs 30.1%, P = 0.006), atrioventricular block (32.7% vs 19.7%, P < 0.001), multi-organ failure (19.7% vs 6.86%, P < 0.001), longer intensive care unit (6 vs 5 days, P < 0.001) and hospital stay (20 vs 15 days, P < 0.001) in combined TVR group over isolated TVR patients. Surprisingly, no difference in long-term survival between these 2 groups was reported (Fig. 4A). Conversely, there is a statistically significant difference in favour of the combined TVR group regarding the incidence of long-term reoperation (Fig. 4B). We believe this finding is influenced by the incidence of endocarditis. In particular, of the 31 patients who underwent reoperation, 15 of them were originally diagnosed with endocarditis and treated with isolated TVR. Twelve of these patients required reoperation due to reinfection during follow-up.

Figure 2E shows that freedom from RV dysfunction was 88.1% at 5 years, 80% at 10 years and 74% at 15 years. We believe that this aspect is important in therapeutic decisions during the follow-up of these patients. As suggested by previous studies [16], patients who were previously treated with TVR had high-risk profiles, with important comorbidities, often associated with RV dysfunction. All these elements discourage both the patient and the surgeon from undertaking surgical reoperation. In this context, valve-in-valve transcatheter strategies are rapidly advancing and could represent a game changer in the treatment of these patients. The benefits of these new technologies include their ease of implantation, reduced invasiveness and the possibility of treating patients who are at high surgical risk.

In conclusion, this study highlights that patients requiring TVR represent a high-risk, frail population, with both isolated and combined surgical procedures being associated with significant early and late mortality. However, the use of bioprosthetic valves has demonstrated long-term efficacy, providing durable freedom from structural valve degeneration for up to 10 years post-implantation.

Limitations

There are some limitations to our study that should be considered. First, the retrospective nature of the study affects both the completeness and the granularity of the data, this could include both clinical and echocardiographic data. Second, we were unable to conduct a detailed analysis focused on each model of bioprosthesis used. Third, although the study involved 10 international institutions, the sample size is still insufficient to allow for a detailed analysis of each type of pathology included.

Last, the number of patients with long follow-up (15 years) was very limited. For this reason, readers should consider 15-year results with caution.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

No funding source was provided for this study.

Conflict of interest: Dr Michael A. Borger discloses that his hospital receives speakers’ honoraria and/or consulting fees on his behalf from Edwards Lifesciences, Medtronic, Abbott, and Artivion. Dr Basel Ramlawi has received grants, personal fees, and nonfinancial support from Medtronic, Liva Nova, and AtriCure. Dr Gianluca Lucchese discloses that he has received speakers’ honoraria and/or consulting fees from Edwards Lifesciences, Bard/BD, Artivion and Abbott. Dr Thomas Modine is a consultant and Senior Advisory Board member for Medtronic and reports grant support and consultant fees from Edwards and Abbott. The remaining authors have no conflicts of interest or financial relationships with the industry to disclose.

DATA AVAILABILITY

The data cannot be shared for privacy reasons.

Author contributions

Antonio Piperata, MD: Conceptualization; Methodology; Writing—original draft; Writing—review & editing. Jef Van den Eynde: Formal analysis; Methodology; Software; Writing—original draft; Writing—review & editing. Mateo Marin-Cuartas: Conceptualization; Data curation; Methodology. Giacomo Bortolussi: Data curation. Petr Fila: Data curation. Tim Walter: Data curation. Mehmet Cahit Sarıcaoğlu: Data curation. Jan Gofus: Data curation. Bilkhu Rajdeep: Data curation. Michel Pompeu Sá: Data curation. Fabrizio Rosati: Data curation. Manuela De La Cuesta: Data curation. Elisa Gastino: Data curation. Besart Cuko: Data curation. Julien Ternacle: Data curation. Carlo de Vincentiis: Data curation; Project administration. Martin Czerny: Data curation; Supervision; Validation. Ahmet Rüçhan Akar: Project administration; Supervision; Validation. Gianluca Lucchese: Data curation; Project administration. Basel Ramlawi: Project administration; Supervision; Validation. Michael Borger: Project administration; Supervision; Validation. Thomas Modine: Conceptualization; Methodology; Supervision; Validation

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Paolo Denti, Nestoras Papadopoulos, Thierry A. Folliguet and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• AATS

  American Association of Thoracic Surgery

 
• CI

  Confidence interval

 
• EACTS

  European Association of Cardio-Thoracic Surgery

 
• HR

  Hazard ratio

 
• NYHA

  New York Heart Association

 
• RV

  Right ventricular

 
• SHR

  Subdistribution hazard ratio.

 
• STS

  Society of Thoracic Surgeons

 
• TR

  Tricuspid regurgitation

 
• TVR

  Tricuspid valve replacement

Author notes