Low Incidence of Structural Valve Degeneration With the Trifecta Aortic Valve Bioprosthesis.

An ideal bioprosthetic heart valve would have excellent haemodynamic performance and a low rate of Structural Valve Degeneration (SVD). The objective of this study was to investigate the haemodynamic performance and incidence of reintervention due to SVD of the Trifecta aortic valve. Methods: Between 2012 to 2019, 2934 valve surgeries were performed in which 2185 were biological valves. Out of 2185 patients, only 399 patients had undergone Trifecta biological Aortic Valve Replacement (AVR). Three surgeons performed all aortic Trifecta valve surgeries. Operative mortality, incidence of reintervention and overall survival gures was calculated. Results: Conclusions:

The most common interventions for valvular heart disease in developed countries are valve replacements for aortic stenosis (4). Valve replacement can either be performed with a Mechanical Heart Valve (MHV) or Biological Heart Valve (BHV). Over recent years there has been a worldwide shift from the use of MHVs to BHVs in the treatment of aortic stenosis (5). In the United Kingdom (UK), bioprosthetic valves have been more commonly used than mechanical valves since 1999, and in 2012 approximately 86% of all aortic valve replacements were performed with a biological valve(6). A similar trend has also been demonstrated in the USA (7).
No prosthetic heart valve is perfect and both MHVs and BHVs have advantages and disadvantages.
MHVs are durable but require life-long anti-coagulation. Although, BHVs do not require treatment with anticoagulation, they are at risk of structural valve degeneration (SVD). The use of BHVs has been shown to be associated with a better safety pro le and sustained lifestyle quality(8, 9). There are numerous BHVs commercially available and long-term clinical outcomes have been shown to vary between devices(6).
The Trifecta GT with anti-calci cation (Abbott™, Illinosis, USA) is a BHV designed for supra-annular placement in the aortic position. Previously published small scale and large-scale single centre studies have demonstrated excellent long-term durability of this Trifecta valve (10)(11)(12)(13)(14). On July 2020, the Medicines and Health Regulatory Authority (MHRA) issued a Medical Device Alert regarding cases of structural valve degeneration on all Trifecta series of valve patients to be followed up (15). The objective of this study was to present our centres experience and results obtained with this prosthesis with followup over 7 years.

Methods:
All patients who underwent aortic valve replacement with Trifecta and Trifecta Glide Technology (St. Jude Medical™ and Abbott™, USA) aortic valves between 2012 and 2019 at a single UK centre were included.
The decision about whether to implant a BHV or MHV was made by each individual patient in conjunction with the operating surgeon taking into consideration the management of valvular heart disease guidelines of the European Society of Cardiology and the European Association for Cardiothoracic surgery (2). The choice about which speci c device was implanted was made by the operating surgeon.
All preoperative, intraoperative, and postoperative clinical data were collected from a prospectively maintained clinical database. All echocardiographic data were obtained from intra-operative transoesophageal echo (TOE) or valve surveillance databases. Survival status was determined from the National Health Service (NHS) Patient Administration Lorenzo survival database.

Isolated AVR technique (sternotomy):
A full median sternotomy is performed, the pericardium is incised and the patient is heparinised. The heart is cannulated and cardiopulmonary bypass commenced. Antegrade and retrograde cardioplegia cannulas are inserted. The aortic cross clamp is applied and cardioplegia administered. Once the heart is adequately protected and vented, the aortic valve is exposed, using either stay sutures or hand-held retractors. Venting is performed either via the right superior pulmonary vein, directly via the aortic root or occasionally via the pulmonary artery. The native valve is excised and the annulus decalci ed. The annulus is then measured and an appropriate prosthesis chosen. Multiple interrupted sutures are placed through the aortic annulus and then through the sewing ring of the valve. The sutures are then tied and cut. The aortotomy is closed as the patient is rewarmed. De-airing manoeuvres are performed and the cross-clamp removed. Ventricular and atrial epicardial pacing wires are attached prior to routine closure.
Isolated AVR technique (minimal access): All minimal access procedures were performed via an upper hemi-sternotomy. The sternum is divided in the midline below the level of the 3 rd intercostal space and then laterally into each intercostal space. The visible pericardium is incised and aortic cannulation is performed as normal with cannulation of the right atrium performed either directly or via the superior vena cava depending on access.
Antegrade cardioplegia is given via the aortic root and subsequently directly using handheld cannulas into the coronary ostia. Venting is performed via the aortic root. Excision of the native valve and implantation of the prosthetic valve is performed as described for full sternotomy. Routine minimally invasive closure techniques were performed.

Outcomes:
The primary outcome for the study was re-intervention for structural valve degeneration.
Echocardiographic outcomes included mean and peak postoperative gradient across the aortic valve. Other secondary outcomes were postoperative complications, length of hospital stay, 30-day mortality and overall 1-year and 3-year survival.

Statistical analysis:
All statistical analysis were performed using Prism v7.01. Descriptive statistics were used to demonstrate patient demographics and are expressed as mean ± standard deviation unless otherwise speci ed. Chisquare tests were performed for categorical variables and Student T tests were performed to compare continuous variables. When there were 3 or more paired time points to be analysed then the repeated measures one-way ANOVA was used followed by Tukey's corrected multiple comparisons tests. Overall survival was evaluated by Kaplan-Meier survival curves. Right censoring of patients occurred when they survived to the end of the study. Survival analyses were performed for the total cohort and strati ed by age (<75 years versus ≥75 years) or by valve size (≤23mm vs. ≥25mm). Group survival comparisons were performed with the Log-rank (Mantel-Cox) test. Statistical signi cance was accepted when p≤0.05.

Results:
Demographics and preoperative variables: A total of 399 procedures met the inclusion criteria and these were performed in 399 individual patients. The majority of patients were male (260, 65.2%) and the overall mean age of the cohort was 73.4 ± 8.0 years. The majority of patients had good pre-operative left ventricular function (281, 70.4%). Surgery was predominantly performed on an elective basis although 112 (28.1%) procedures were non-elective. There were 22 patients (5.5%) who had undergone previous cardiac surgery. The mean (±SD) logistic EuroSCORE was 7.4 (±2.7). Full pre-operative patient characteristics for the cohort are displayed in Table  1.
Intraoperative outcomes: Over 90% of procedures were performed in patients with moderate or severe aortic stenosis and 16.3% of patients had moderate or severe aortic regurgitation. Most patients underwent either isolated AVR (50.6%) or in combination with Coronary Artery Bypass Grafting (CABG) (40.1%), CABG + valve + other procedures (4%) and valve replacement with other procedures (5.3%). Other procedures included concomitant aortic surgery, additional valve surgery or surgical treatment of atrial brillation. There were 14% procedures performed via a minimally invasive approach and all of these were isolated AVR. Cumulative bypass time mean was 134.0 ± 49.0 and cumulative cross clamp mean time was 104.4 ± 38.1. The most common valve size implanted was 23mm. Full intraoperative details are displayed in Table 2.

Reintervention and survival:
The majority (93.5%) of patients have not required any type of repeat surgery and one patient (0.3%) required re-intervention for structural valve degeneration. Survival to 30 days was 97.2% with 11 patients having died during this time. The Kaplan Meier plot for overall survival is displayed in Figure 1. Of the patients who had their surgery at least one year prior to the end of the study (n=308), 287 survived to 1 year (94.1%). Of the patients whose surgery occurred at least 3 years before the end of the study (n=162), there are 129 survivors (79.6%). When strati ed by age, mortality up to 5 years follow-up was not signi cantly different between the groups (<75 years vs. ≥75 years, p=0.11, gure 2). Furthermore, no differences are observed in survival when strati ed for valve size (valves ≤23mm vs. valves ≥25mm, p=0.67, gure 3) Valve gradients: As shown in table 4, there was no incidence of severe paravalvular or transvalvular regurgitation in patients who have undergone echocardiographic surveillance at either 1-year (n=235) or 3 years (n=56).

Discussion:
This single centre retrospective analysis of Trifecta valve utilisation demonstrates post-operative complication rates in line with other series (9,16,17). The incidence of aortic valve reoperation due to structural valve degeneration was low, with only a single patient requiring repeat aortic valve replacement at 3 years post-surgery for transvalvular leak due to a non-calci c lea et tear. This was a 34-year-old man who had a previous AVR 10 years earlier with another BHV and wanted to avoid anticoagulation due to his lifestyle preferred to have another BHV. There are a range of BHVs available, but it can be either stented or stentless. Stentless BHVs were developed in an attempt to minimise the transvalvular gradient and improve the effective ori ce area (11,17). However, recent studies have suggested that stentless porcine valves have reduced durability (11) and an increased rate of SVD requiring re-intervention(18).
Most BHVs used in contemporary practice are now stented and various bovine pericardial valve prosthesis have demonstrated initially better long-term durability and excellent haemodynamic performance (19). However, few studies have demonstrated that there is a high chance of SVD on trifecta(10, 12, 16) and pericardial mitral ow (19,20) aortic prosthetic valve post-surgery.
The Trifecta valve is a stented pericardial bovine heart valve designed for supra-annular placement. It was rst introduced for patient use in 2007. The valve lea ets are externally mounted on the stent which allows cylindrical opening during systolic contraction to provide an enhanced effective valve ori ce area and optimised transvalvular gradients. Several literature reports identi ed that rst-generation Trifecta aortic valve was at risk of cusp tears or lea et calci cation (21)(22)(23)(24). The latest generation Trifecta GT valve was introduced in 2016 to provide added handling protection for the stent and lea ets (14). Modi cations included a streamlined and enhanced holder, a softer sewing ring, an additional titanium band in the stent base, optimised lea et suturing techniques and collagen alignment.
The rate of SVD in our study is lower than other previously published studies that have demonstrated a 1% incidence of repeat operation with Trifecta valve for SVD at 5 years(10), and 3.3% at 7 years(16).
Studies have also reported that younger patients have a higher incidence of SVD (3 years: 2.4%, 5 years: 4.2% and at 7 years: 27.9%)(22). This was the case in our study with our single patient who is 34 years old, exhibiting SVD. Previous studies of the trifecta valve have reported the incidence of device related stroke, systemic embolism of 2.7-6.0% (12). In our cohort we observed a post-operative stroke rate of 0.8% in a patient who underwent combined AVR + CABG with no systemic embolic complications.
Valve performance in our cohort was good and in 3 year follow up data, there was no patient who exhibited severe paravalvular regurgitation and only one patient who had developed a moderate paravalvular regurgitation. The survival rate at hospital discharge was 96.7%, at 30 days was 97.2%, at 1 year 94.1% and at 3 years was 79.6%. Irrespective of having an older population in our cohort of patients, the survival rate was better than other studies who had survival rates ranging from 70-88.7% (11,14,16). The incidence of a mild paravalvular leak at one year was 2.6% (n = 287 patients at risk) and minimal transvalvular leak was 21.7%. Haemodynamic performance was very favourable in the case series with mean valve gradients at one year of 9.5 ± 4.1 mmHg compared to ranges from 10.1 to 20mmHg in the published literature (10,11,14).
Our experience demonstrates that the use of the Trifecta valve is safe and achieves good clinical outcomes. However, like any bioprosthetic valve it has its own advantages and disadvantages. In our experience avoiding over-sizing of the valve, careful knot tying and valve handling to avoid inadvertent damage to the lea ets are the key elements to achieving good results. We suggest that unnecessary handling of lea ets at the time of the implant may be a signi cant contributing factor to the early structural valve deterioration that has been reported in other studies(16, 22). We believe that proctoring and adequate training when a surgeon rst uses the valve are vital to ensuring good results. We will also urge surgeons new to the use of this valve to exercise caution if they implant the valve with the Cor-Knot® system to avoid deploying the Cor-Knot® in such a way that it may be in contact with the valve lea et and initiate SVD. In our practice, we try to avoid using Cor-knot especially if we are implanting the Trifecta valve.

Limitations:
There are a number of limitations of this study. Firstly, it is a retrospective analysis of prospectively collected data with a small sample size from a single centre and as such the generalisability of these results to other centres is unknown. A further limitation is that echocardiographic follow-up was not available for all patients at either one or three year. This is due to the retrospective nature of the study and variations in regional surveillance programmes. 1-year echocardiographic data was available for two thirds of the original cohort and as such the results at this time are likely to be robust. However, 3-year follow-up echocardiographic information was only available for 13.5% of the original cohort. Despite this however it is encouraging that there has been no signi cant change in the gradients across the valves in these patients. Development of mild transvalvular regurgitation in around one quarter of patients between one and three years is potentially concerning, and these patients should continue to be followed up on a long-term basis. It should be noted that this study includes data from two versions of the Trifecta valve. Although the Trifecta GT contains a number of modi cations the fundamental design of the valve is the same and as such no comparisons have been made between the two valve types. Another limitation is that this study represents a heterogeneous group of procedures. Although the key message regarding the safety pro le of the Trifecta valve is not affected by this issue, it does mean that direct comparisons of the results from this study with other more homogenous groups should be undertaken with caution.
Long-term multicentre data is required to ensure the on-going safety issues of any implantable medical device. Unfortunately, the UK Heart Valve Registry was defunded and as such prospective and retrospective systematic monitoring of heart valve performance is not currently possible in UK. Ideally, as is the case with other devices such as hip replacements, prospective national registries with serial echocardiographic data for all biological valves implanted should be established and would provide an invaluable resource. Establishing such registries to facilitate long-term performance monitoring is now even more important given the rapid expansion in utilisation of transcatheter aortic valve implantation.

Conclusion:
This study of patients who have received a Trifecta bioprosthetic aortic valve demonstrates good short and mid-term outcomes out to 3 -years with a low rate of structural valve degeneration requiring re- Open Access funding will be enabled and organised from the Abbott company or from the heart charity ( *Please note that this will be updated near the submission time) Availability of data and materials: Please contact corresponding author for data request.
Ethical approval and participant consent: This retrospective study complied with the Helsinki Declaration (2000) and approval to perform this analysis was given by the local ethical governance committee (registration no: 9711), based on retrospective data retrieval. For this reason, the ethical committee waived the need for written patients' informed consent.

Consent for publication:
Yes, we have permission to publish this article and results of this study. All authors read and given permission to publish this article.   Valve ratio (mean ± SD) 0.5 ± 0.1 Figure 1 This gure illustrates the overall survival of the cohort to 3 years.

Figure 2
This gure illustrates the peak and mean valve gradients at 1 and 3 years for all patients with 3-year follow-up data available (n=54).

Figure 3
This gure illustrates the survival strati ed by valve size.