Mid-term clinical and haemodynamic results after aortic valve replacement with the Trifecta bioprosthesis

Abstract

 

OBJECTIVES

The aim of this study was to evaluate the clinical and haemodynamic results after implantation of the Trifecta bioprosthesis.

METHODS

This study is a retrospective analysis of all patients undergoing Trifecta aortic valve replacement between 01 January 2012 and 31 December 2017 at the Ghent University Hospital. Univariable and multivariable analyses were performed to identify predictors of valve- and procedure-related complications and mortality. The haemodynamic performance was analysed by longitudinal Doppler echocardiography.

RESULTS

The mean age of the 182 patients was 77 [standard deviation (SD): 5.5] years; 54.9% were women. The mean follow-up was 39.8 (SD: 24.3) months. Overall survival at 1 and 5 years was 86% (SD: 3%) and 68% (SD: 4%), respectively, and overall freedom from structural valve deterioration was 100% and 98% at 1 and 5 years, respectively. There was no valve thrombosis nor early endocarditis. Urgent surgery was the only risk factor for early mortality in the multivariable analysis [P = 0.009, odds ratio 0.06, 95% confidence interval (CI) 0.01–0.5]. Preoperative atrial fibrillation was the most important predictor of late mortality (P = 0.001, hazard ratio 3.68, 95% CI 1.65–8.21). The average peak gradients were stable from discharge up to 1 and 5 years postoperatively [15 (SD: 6) and 17 (SD: 8) mmHg].

CONCLUSIONS

These results confirm the excellent clinical performance of the Trifecta valve, particularly in an elderly age group. Through the 7-year follow-up period, low transvalvular gradients persisted, and only a few patients needed reoperation. Although structural valve degeneration occurred rarely, it was unrelated to valve size or age at implantation; therefore, further long-term follow-up remains mandatory.

Currently, the majority of prosthetic valves used for surgical aortic valve replacement (SAVR) are bioprosthetic valves.

INTRODUCTION

Currently, the majority of prosthetic valves used for surgical aortic valve replacement (SAVR) are bioprosthetic valves. Superior haemodynamics and reliable durability are the main criteria for prosthetic valve selection, to compete with the increasing tendency for transcatheter aortic valve implantation transcatheter aortic valve replacement (TAVR) [1–3].

Several bioprosthetic valves are on the market nowadays, each with their own specific characteristics. The Trifecta bioprosthesis (Abbott) is a 3-leaflet stented pericardial valve developed for supra-annular aortic placement. The design of the Trifecta valve has specific characteristics to solve some of the fundamental limitations of other aortic valve bioprostheses. The scalloped sewing ring is concave to provide an adequate clearance of the native coronary arteries. A sheet of bovine pericardium is wrapped externally around the titanium stent, which allows a maximal cylindrical systolic opening, resulting in lower gradients and higher effective orifice area than any other stented aortic prosthesis. This favourable haemodynamic performance leads to a very low incidence of severe patient–prosthesis mismatch (PPM) [4]. Furthermore, the intrinsic distensibility of the titanium stent allows further expansion of the prosthesis in high-pressure loading conditions [4].

Despite the excellent haemodynamic performance of the Trifecta prosthesis, recent reports have suggested a more cautious approach in terms of its durability. The Trifecta bioprosthesis appears to be associated with a greater rate of structural valve deterioration (SVD) and a higher need for repeat aortic valve replacement for SVD, than some other bioprostheses [5–8].

The aim of this study was to investigate the clinical and haemodynamic performance of the Trifecta bioprosthesis in a single-centre setting and to compare the results with the currently available Trifecta valve data, as well as with the performance of other aortic valve bioprostheses.

PATIENTS AND METHODS

Ethical statement

The study was approved by the Ethical Committee of Ghent University Hospital (nr. B670201836876–B670201838877—10 July 2018). Due to the retrospective nature of the study, patient consent was waived.

Study design and methodology

All consecutive patients who underwent isolated or combined SAVR with the original model Trifecta aortic valve prosthesis at the Cardiac Center of the Ghent University Hospital, Belgium, between January 2012 and December 2017 were included (Fig. 1). In patients >70 years of age, and with a small annulus for size or risk for PPM, the departmental policy is to choose between a Trifecta or a sutureless Perceval bioprosthesis. Patients were followed annually up to 8 years postoperatively, both in referral community hospitals as in the Ghent University Hospital.

Figure 1:

Flowchart of surgical aortic valve replacement in Ghent University Hospital.

Surgery was performed mainly by median sternotomy, and the prosthesis was routinely implanted supra-annularly with single non-pledgetted sutures.

The electronic hospital database was searched for clinical and haemodynamic parameters (Supplementary Material, Appendix SI). Some parameters were defined by guidelines. Valve-related parameters were reported according to the guidelines of Akins et al. [9] (Supplementary Material, Appendix SI). Follow-up data were obtained from outpatient clinic visit reports, containing clinical and echocardiographic data. The closing date for follow-up was set on 8 July 2020.

The primary end point of the study was in-hospital mortality. Early adverse events were defined as valve-related and non-valve-related events developing within 30 days post-implant or during the index hospitalization. Secondary end points were freedom from all-cause mortality and the occurrence of late valve-related and non-valve-related adverse events (Supplementary Material, Appendix SI). The echocardiographic haemodynamic parameters during follow-up were collected at different time points.

Statistical analysis

The statistical analysis was performed with SPSS Statistics 26.0 Statistics Software (IBM Corporation, Armonk, NY, USA) and R version 3.5.3 (R Foundation for statistical analysis, Vienna, Austria). Categorical variables are presented as numbers and percentages. Continuous variables are presented as mean and standard deviation (SD), or median and interquartile range when not normally distributed. The Shapiro–Wilk test was used to explore for normality. Continuous characteristics were compared with the Kruskal–Wallis test to compare the aortic valve peak gradients between the different valve sizes, and the Friedman test to detect significant changes in the valve gradients during follow-up. Univariable analysis, using univariable binary logistic regression, was performed to identify relationships between early and late mortality and complications, and patient or perioperative factors. The variables taken into account for the univariable analysis are indicated in italic in Supplementary Material, Appendix SI. Variables with a P-value of <0.2 were entered into a multivariable logistic regression analysis (backward conditional stepwise method) to determine independent influence on death and complications. Collinearity diagnostic tests were used to exclude variables that could lead to multicollinearity. The effect of significant determinants was expressed as the odds ratio with 95% confidence intervals (CIs). Survival was analysed using the Kaplan–Meier survival analysis. The log-rank (Mantel Cox) test was performed to find out significant differences between groups. Significant factors were entered into a multivariable Cox-regression analysis model, and given by the hazard ratio and 95% CI. The proportional hazard assumption was verified graphically with the log–log plot as well as by the Schoenfeld residuals. A result was considered significant for a two-tailed P-value of <0.05. Significance in the Kaplan–Meier survival analysis was considered valid up to 6 years of follow-up to maintain at least 10% of the population at risk.

RESULTS

Clinical results

The preoperative characteristics of the 182 patients are summarized in Table 1. At the date of surgery, the majority of the patients was older than 70 years (n = 165, 91%). Operative characteristics are summarized in Table 2. A 23-mm valve was most frequently used (40.1%), followed by a 21-mm valve (30.2%). There was no implantation of a 29-mm Trifecta valve.

Table 1:

Baseline preoperative characteristics

Variables	Study population (n = 182)
Male gender	82 (45.1)
Age (years)	77 (SD: 5.5)
Weight (kg)	75 (SD: 13.0)
Height (cm)	165 (SD: 9.0)
Body mass index (kg/m2)	27.6 (SD: 4.6)
Body surface area (m2)	1.8 (SD: 0.2)
Comorbidities
Hypertension	138 (75.8)
Dyslipidaemia	118 (64.8)
Diabetes mellitus	46 (25.3)
Active smoking	15 (8.2)
Peripheral arterial disease	65 (35.7)
Chronic kidney disease	15 (8.2)
COPD	24 (13.2)
Prior cerebrovascular disease	24 (13.2)
Prior myocardial infarction	25 (13.7)
Medication
Antihypertensive drugs	104 (57.1)
Betablockers	83 (45.6)
Cholesterol lowering drugs	118 (64.8)
Diuretics	79 (43.4)
Anticoagulants (VKA/NOAC/heparin)	46 (25.3)
Antiplatelet drugs	102 (56.0)
Prior cardiac surgery	17 (9.3)
Coronary artery bypass graft	9 (4.9)
Aortic valve replacement	8 (4.4)
Mitral valve replacement	1 (0.5)
Tricuspid valve repair	1 (0.5)
Atrial septal defect repair	1 (0.5)
Preoperative cardiac rhythm
Sinus	144 (79.1)
Atrial fibrillation	32 (17.7)
Pacemaker	6 (3.3)
Degree of urgency
Elective	162 (89.0)
Urgent	20 (11.0)
NYHA classification
I	22 (12.1)
II	91 (50.0)
III	50 (27.5)
IV	19 (10.4)
EuroSCORE II	5.7 (SD: 9.8)
Primary valve disease
Aortic stenosis	174 (95.6)
Aortic regurgitation	7 (3.8)
Prosthetic valve endocarditis	1 (0.6)

Continuous variables are presented as the mean (SD) and categorical variables as n (%).

COPD: chronic obstructive pulmonary disease; EuroSCORE: European system for cardiac operative risk evaluation; NOAC: novel oral anticoagulants; NYHA: New York Heart Association; SD: standard deviation; VKA: vitamin K antagonists.

Table 2:

Operative data

Variables	Study population (n = 182)
Cardiopulmonary bypass time (min)	108 (SD: 40)
Isolated aortic valve replacement	83 (SD: 24)
Combined procedure(s)	125 (SD: 39)
Aortic cross-clamp time (min)	76 (SD : 28)
Isolated aortic valve replacement	60 (SD: 17)
Combined procedure(s)	87 (SD: 29)
Implanted prosthesis size (mm)
19	11 (6.0)
21	55 (30.2)
23	73 (40.1)
25	36 (19.8)
27	7 (3.8)
Concomitant procedurea	108 (59.3)
Coronary artery bypass graft	72 (39.6)
Mitral valve repair	24 (13.2)
Mitral valve replacement	9 (4.9)
Tricuspid valve repair	17 (9.3)
Ascending aortic replacement	5 (2.7)
Aortic root enlargement	6 (3.3)
Aortic root reconstruction	4 (2.2)
Bentall procedure	1 (0.5)
Maze procedure	6 (3.3)
Patent foramen ovale closure	1 (0.5)
Left atrial appendage occlusion	6 (3.3)
Septal myectomy	3 (1.6)

Continuous variables are presented as the mean (SD) and categorical variables as n (%).

^aMore than 1 concomitant procedure was possible.

SD: standard deviation.

Early adverse events

Early adverse events are summarized in Table 3. Hospital mortality occurred in 16 patients (8.8%): 6.7% (5/74) in isolated SAVR, 4.4% (2/46) in SAVR with coronary artery bypass graft only and 14.5% (9/62) in SAVR with other combined procedures. Cardiac failure and infections were the main causes of early mortality. With regard to the valve-related complications, there were 7 early thromboembolic events, including 6 strokes (3.3%) and 1 gastrointestinal embolic event (0.5%). Of these 7 patients, 2 were known to have persistent atrial fibrillation. There were no instances of early valve thrombosis, endocarditis, SVD or valve explant. Postoperative pacemaker implantation was needed in 2 patients (1.1%). The most important procedure-related adverse event was a major bleeding with need for revision (7.1%).

Table 3:

Complications

Variables	Early complications (n = 182)	Late complications (n = 166)	Late complications/ patient-years (%)
Valve-related events
Thromboembolic events	7 (3.8)	13 (7.8)	2.15
Stroke	6 (3.3)	6 (3.6)
TIA		4 (2.4)
Gastrointestinal	1 (0.5)
Cardiac		2 (1.2)
Pulmonary		1 (0.6)
Prosthetic valve thrombosis	0	0
Non-structural deterioration	3 (1.6)	6 (3.6)
Paravalvular leak	3 (1.6)	3 (1.8)
Other	0	3 (1.8)
Structural deterioration		3 (1.8)	0.5
Endocarditis	0	2 (1.2)	0.33
Arrhythmia requiring pacemaker	2 (1.1)	14 (8.4)
Aortic valve reoperation		4 (2.4)
Procedure-related events
Major bleeding (requiring revision)	13 (7.1)
Revision for bleeding	9 (4.9)
Exploratory sternotomy	1 (0.5)
Secondary sternal closure	1 (0.5)
Reoperation
Mitral valve repair	1 (0.5)
Cardiac tamponade	4 (2.2)
Mortality	16 (8.8)a	42 (25.3)
MOF	2 (1.1)	7 (4.2)
Infection	4 (2.2)	4 (2.4)
Neoplasm		7 (4.2)
Endocarditis		1 (0.6)
Cardiac	8 (4.4)
Neurological	1 (0.5)
Unknown	1 (0.5)	23 (13.9)

Categorical variables are presented as n (%).

^aIn-hospital mortality.

MOF: multiple organ failure; TIA: transient ischaemic attack.

Late adverse events

The mean duration of follow-up was 39.8 (SD: 24.3) months [median 41 (21–60) months] and was 91.2% complete, including 604 patient-years. Late adverse events are summarized in Table 3. The late mortality was 25.3%; in half of the deaths, the cause was unknown. The Kaplan–Meier estimated freedom from all-cause mortality is presented in Fig. 2. Overall survival at 1 and 5 years was 86% (SD: 3%) and 68% (SD: 4%), respectively. The majority of the 13 (7.8%) late thromboembolic events were strokes (n = 6), and of the 13 patients, 5 were known to have atrial fibrillation. There were 2 cases of endocarditis (1.2%) at 5 and 53 months, respectively: 1 was reoperated and 1 died. Non-structural valve dysfunction occurred in 6 patients (3.6%), including 3 paravalvular leaks (1.8%), 1 Dressler syndrome (0.6%), 1 valvular cardiomyopathy (0.6%) and 1 heart failure (0.6%). There were 3 instances of SVD (1.8%), all presenting with severe calcific stenosis at 55, 69 and 80 months after implantation, all in patients >75 years at the index operation. At 6 years, 58% of patients were alive and free from SVD, and the cumulative probability of SVD with death as a competing risk was 3%. The 1- and 5-year freedom from SVD and from reoperation for SVD was 100% (SD: 0%) and 98% (SD: 2%), respectively (Fig. 3A). There were 4 valve reoperations (2.4%): 2 TAVR for SVD (1.2%) and 2 SAVR for endocarditis (0.6%) and for a major paravalvular leak (0.6%). The 1- and 5-year freedom from aortic valve reoperation was 100% (SD: 0%) and 96% (SD: 2%), respectively (Fig. 3B).

Figure 2:

Kaplan–Meier survival analysis after surgical aortic valve replacement with the Trifecta aortic valve prosthesis: overall survival.

Figure 3:

(A) Kaplan–Meier survival analysis: freedom from structural valve dysfunction. (B) Kaplan–Meier survival analysis: freedom from all-cause reoperation. SVD: structural valve deterioration.

Predictors of complications and mortality

The results of the univariable and multivariable analyses of in-hospital mortality and late mortality are summarized in Table 4. The results of the univariable and multivariable analyses of early and late complications are displayed in Supplementary Material, Table S1. In the univariable analysis, the most significant predictors of early complications and in-hospital mortality were urgent surgery, prior cardiac surgery, high European system for cardiac operative risk evaluation (EuroSCORE) II, high preoperative New York Heart Association (NYHA) classification, preoperative atrial fibrillation, chronic obstructive pulmonary disease, decreased postoperative left ventricular ejection fraction, concomitant mitral valve procedure, chronic kidney disease and prior myocardial infarction.

Table 4:

Predictors of mortality

Variablesa	In-hospital mortality				Late mortality
	Univariable P-value	OR (95% CI)	Multivariable P-value	OR (95% CI)	Univariable P-value	OR (95% CI)	Multivariable P-valueb	Hazard ratio (95% CI)
Urgent procedure	<0.001	12.05 (4.17–33.33)	0.009	16.67 (2–100)	0.002	7.14 (2–25)	0.011	4.55 (1.43–14.29)
Prior cardiac surgery	<0.001	8.33 (2.56–25)	0.069	5.26 (0.88–33.33)
EuroSCORE II	<0.001	1.11 (1.05–1.17)			0.002	1.12 (1.04–1.21)
Preoperative LVEF	0.063	2.7 (0.94–7.69)	0.865	1.15 (0.25–5.26)	0.018	2.56 (1.18–5.56)	0.293	1.47 (0.72–2.94)
Postoperative LVEF	0.041	10 (1.1–100)
Preoperative NYHA classificationc	0.008		0.871		0.07		0.472
Preoperative atrial fibrillation	0.005	4.76 (1.58–14.31)	0.092	4.19 (0.79–22.11)	0.005	3.56 (1.47–8.65)	0.001	3.68 (1.65–8.21)
Concomitant CABG					0.044	2.08 (1.02–4.17)	0.256	1.45 (0.76–2.78)
Concomitant tricuspid valve repair	0.034	3.85 (1.11–14.29)	0.57	1.85 (0.22–16.67)
Concomitant mitral valve replacement	0.017	6.25 (1.37–25)	0.419	2.7 (0.24–33.33)
Concomitant mitral valve repair	0.155	2.45 (0.71–8.33)	0.947	1.06 (0.17–6.67)
Chronic kidney disease	0.018	4.76 (1.3–16.67)	0.706	1.47 (0.2–11.11)	0.123	2.63 (0.77–9.1)	0.739	1.19 (0.42–3.45)
Hypertension					0.032	3.03 (1.1–8.33)	0.03	2.94 (1.11–7.69)
COPD	0.034	3.57 (1.1–11.11)	0.849	1.19 (0.19–7.69)	0.505	1.43 (0.51–4)
Diabetes mellitus	0.245	1.89 (0.65–5.56)
Peripheral arterial disease					0.076	1.92 (0.93–3.85)	0.374	1.35 (0.7–2.63)
Prior cerebrovascular disease	0.155	2.44 (0.71–8.33)	0.135	3.33 (0.68–16.67)	0.01	3.57 (1.37–9.1)	0.086	2 (0.91–4.35)
Prior myocardial infarction	0.042	3.33 (1.04–10)	0.196	3.13 (0.56–16.67)

Variables included in the multivariable regression model for in-hospital mortality: preoperative LVEF, preoperative NYHA classification, concomitant mitral valve replacement, concomitant mitral valve repair, concomitant tricuspid valve repair, COPD, chronic kidney disease, prior cardiac surgery, urgent procedure, prior AMI, preoperative cardiac rhythm and prior cerebrovascular disease. Variables included in the multivariable regression model for late mortality: preoperative LVEF, preoperative NYHA classification, concomitant CABG, chronic kidney disease, urgent procedure, preoperative cardiac rhythm, arterial hypertension, prior cerebrovascular disease and peripheral arterial disease. The EuroSCORE II was not included in the multivariable analyses due to collinearity. The boldface values are the significant ones.

^aOnly the most significant predictors of the univariable analyses are displayed.

^b P-value of the multivariable Cox-regression analysis model.

^cThe P-value is the general P-value of preoperative NYHA classification. There is no corresponding OR (95% CI) for this P-value.

AMI: •••; CI: confidence interval; CABG: coronary artery bypass graft; COPD: chronic obstructive pulmonary disease; EuroSCORE: European system for cardiac operative risk evaluation; LVEF: left ventricular ejection fraction; NYHA: New York Heart Association; OR: odds ratio.

At multivariable analysis, prior myocardial infarction and urgent procedure were significant predictors of early complications and in-hospital mortality, respectively. Significant predictors of late mortality were urgent surgery, atrial fibrillation and hypertension. The impact of the preoperative cardiac rhythm on the all-cause mortality is displayed in Fig. 4.

Figure 4:

Log-rank comparisons for survival according to cardiac rhythm. *P-value for difference in survival between patients with sinus rhythm and patients with atrial fibrillation according to the log-rank (Mantel Cox) test.

Haemodynamic results

The haemodynamic results over time are summarized in Supplemental Material, Table S2. Across all valve sizes the peak aortic valve gradients dropped from 68 (SD: 26) mmHg preoperatively to 15 (SD: 6) mmHg at discharge. The aortic valve gradients did not increase significantly during follow-up (P = 0.258). The peak gradients increased from 17 (SD: 8) mmHg at 1 year to 20 (SD: 15) mmHg at 6 years. Larger valve sizes only resulted in lower aortic valve peak gradients at discharge (P < 0.001), as represented in Supplementary Material, Fig. S1. The echocardiographic left ventricular parameters remained stable during follow-up (Supplementary Material, Table S2).

DISCUSSION

The Trifecta valve was designed to address some of the limitations of other aortic bioprostheses. The externally mounted pericardial sheet on a titanium stent was conceived to increase the effective valve opening in order to improve the haemodynamic performance, meant to be beneficial in patients with short stature and a small aortic annulus, at risk for PPM [10].

The purpose of the current study was to confirm the safety, haemodynamic performance and durability of the Trifecta valve in a population of very elderly patients (mean age 77 years), and a relatively low body surface area (1.8 m²), the majority being females. In our hospital practice, the use of the Trifecta valve represented ∼28% of all bioprosthetic aortic valves implanted surgically over the study period. A surgical bias was undoubtedly present, as the valve choice was partially driven by the promising haemodynamic results published on this valve, demonstrating very low gradients even in small valve sizes, with a low prevalence of PPM [11–15].

The early postoperative outcome was comparable with prior studies on short-term outcomes with this valve [4, 6, 11, 13], and urgent surgery was the only significant predictor of in-hospital mortality. Combined valve procedures were predictive of early events, while coronary artery disease increased the risk for late mortality, as noted by others [11, 16]. Early and late mortality were not valve related. The observed survival we found at 1 and 5 years is a reflection of the high-risk patient profile: a high age at operation, a large number of comorbidities, with 18% of patients presenting with preoperative atrial fibrillation, which turned out to be the most significant risk factor for late mortality. Thromboembolic events occurred at a rate of 2.15% per patient-years, which was comparable to the 0.9% and 1.9% per patient-years as reported by Goldman et al. and Bavaria et al. [4, 13]. Prosthetic valve endocarditis occurred rarely and only late (0.33% per patient-years), and there were no instances of valve thrombosis. Paravalvular leaks were usually minor and disappeared at follow-up, only one patient required a reoperation for a major paravalvular leak.

Our findings confirmed the good haemodynamic performance of this valve as reported previously: very low mean gradients at discharge that are maintained through 7 years of follow-up. Peak gradients inversely correlated with valve size but remained favourable in the smaller diameters as well. At 1 year after implantation, the mean gradients of the Trifecta valve were on average 9 (SD: 7) mmHg, which was equivalent to the gradients of the Sapien and CoreValve TAVR [17]. Looking particularly at the smallest valve sizes, the transvalvular gradients in the Trifecta valve [11 (SD: 3) and 9 (SD: 3) mmHg for the 19- and 21-mm valve, respectively] were clearly superior to the reported mean gradients of 16–21 mmHg across other stented biological valves of 19- and 21-mm size [12, 18–22]. When our results are compared with published haemodynamic profiles of sutureless valves such as the Intuity or Perceval, we obtained slightly lower gradients at discharge, which levelled out and became similar at 2 years of follow-up [23]. The excellent haemodynamic performance of the Trifecta valve reduces the likelihood of severe PPM, known to be associated with increased late mortality and SVD.

In the current study, the prevalence of SVD was 1.8%, and the 5-year reoperation rate for SVD was 0.6%, with an overall 1- and 5-year freedom from reoperation for SVD of 100% and 98%, respectively. A similar 5-year freedom from redo for SVD of 99% and 97,3% was reported by Kilic et al. and Goldman et al. [11, 13]. As our study enrolled very elderly patients, results of SVD or other mechanisms of late valve failure may have been confounded through the high mortality during follow-up. The 3 patients that had a proven SVD with mainly calcific stenosis were all over 75 years of age at implantation, received different valve sizes, and the increased gradients were noted on echocardiography at 55, 69 and 80 months of follow-up. Several reports have emerged on a distinctly higher cumulative incidence of SVD in Trifecta valves compared to other biological stented valves, with failure rates up to 13% at 7 years in the study of Fukuhara et al. [6, 8]. The SVD was found to be related to age at implantation in some studies [6], but this was not a consistent finding. We had no incidences of more than moderate aortic valve incompetence or leaflet tears, often described as the driving mechanism of premature Trifecta valve failure, appearing as early as 3 months postoperatively [24]. In the early cases of failure, the main finding was intraprosthetic regurgitation due to a leaflet tear starting at a stent post, and extending towards the base of the leaflet, without calcification [25–28]. At reoperation, some authors reported a slight deformation of one of the struts, leading to a possible distortion of the valve [29]. The Trifecta valve is indeed quite fragile and delicate, and care must be taken not to damage the valve cusps when tying the knots in a small aortic root. The company has tried to overcome this weakness by recommending to leave the protective holder onto the stent, but often this has to be removed to increase the working space at the base of the valve stent. A small degree of stent distortion, induced at implantation, may elicit a mild central incompetence, not sufficient to lead to immediate serious valve dysfunction, but it may have an untimely effect on preliminary degeneration of the prosthesis.

Limitations

This study has the inherent features of a retrospective and single-centre trial. The prosthesis selection may have been biased by the surgeon’s choice. The serial follow-up echocardiographic studies were not always performed in a standardized way. Echocardiographic follow-up was only 91% complete, as some patients stayed in a nursing home without further echocardiographic investigations. Furthermore, the generalizability of our results needs to be nuanced because of the decreasing sample sizes during the follow-up, due to mortality and missing values, as the number of patients surviving with a useful echocardiographic follow-up available at 7 or 8 years was relatively small. PPM could not be calculated in the majority of the patients, due to missing effective orifice area data. The possibility of uncontrolled biases in the assessment of the haemodynamic performance can also not be ruled out due to the absence of a subgroup analysis between the patients with isolated SAVR and those with a concomitant procedure.

The comparison of the SVD rates between studies is difficult because of the variability in definitions used for SVD and in surgical experience with the implantation of the Trifecta valve.

Despite the non-normally distribution of some variables, the transvalvular gradients are presented as mean (SD) instead of median and interquartile range, to allow comparison with other studies.

CONCLUSION

This retrospective study confirmed the excellent haemodynamic performance of the Trifecta bioprosthesis, with low transvalvular gradients remaining stable up to 7 years of follow-up in an older population at increased risk for PPM. The early complication rates were comparable to prior reports and were mainly patient-related. Although the incidence of structural valve degeneration was quite low in our series, it seemed to be unrelated to valve size or age at implantation, casting some doubt on the exact mechanism of failure. Further investigations with larger patient populations are needed to establish the durability of the Trifecta valve and to explore the pathophysiology of intrinsic failure.

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

ABBREVIATIONS

CI

Confidence interval

PPM

Patient-prosthesis mismatch

SAVR

Surgical aortic valve replacement

SVD

Structural valve deterioration

SD

Standard deviation