Long-term durability after surgical aortic valve replacement with the Trifecta and the Intuity valve—a comparative analysis

Abstract

 

Open in new tabDownload slide

OBJECTIVES

Long-term durability of surgical bio-prostheses is a key factor, especially in the era of transcatheter aortic valve replacement. We compared the incidence of structural valve deterioration (SVD) between patients undergoing surgical aortic valve replacement (SAVR) with the Trifecta (Abbott Laboratories, Abbott Park, IL, USA) or the Intuity valve (Edwards Lifesciences, Irvine, CA, USA).

METHODS

Between April 2010 and May 2020, 1118 patients underwent SAVR with the Trifecta (n = 346) and the Intuity (n = 772) valve at a single centre. A total of 1070 patients (Trifecta n = 298, Intuity n = 772) were analysed after the exclusion of patients with pure regurgitation and endocarditis. Retro- and prospective echocardiographic and clinical follow-up was performed. Cox proportional hazards regression models were performed to identify prognostic factors for SVD, aortic re-interventions and mortality.

RESULTS

With 27 cases (Trifecta n = 23, Intuity n = 4) of SVD observed, cumulative incidence of SVD was significantly higher in the Trifecta cohort (P < 0.001). Implantation of a Trifecta valve [hazard ratio (HR) 11.20; 95% confidence interval 3.79–33.09], log-transformed preoperative creatinine (HR 2.47; 1.37–4.44) and sex (male HR 0.42; 0.19–0.92) emerged as prognostic factors of SVD. A significantly higher cumulative incidence of re-interventions was observed in the Trifecta cohort (P = 0.004) and valve type was an independent time-varying risk factor (HR at 12 months 2.78; 95% confidence interval 1.42–5.45). Overall, no significant differences in all-cause mortality were observed between the groups (log-rank test: P = 0.052).

CONCLUSIONS

SVD was significantly more frequent in patients receiving a Trifecta valve and its implantation was an independent risk factor for the occurrence of SVD and aortic valve re-interventions. This comparative analysis of 2 low-gradient bioprosthesis put the long-term durability of the Trifecta valve in question and need to be taken into consideration when performing bioprosthetic SAVR.

INTRODUCTION

Within the last decades, an evident trend toward an increased use of bioprosthetic valves in patients undergoing surgical aortic valve replacement (SAVR) has been observed [1]. Continuous efforts have been made to improve bioprosthetic valve design and to optimize their haemodynamic performance. Nevertheless, structural valve deterioration (SVD) remains the Achilles’ heel of bioprosthetic valves and long-term durability is a key factor, especially in the current era of transcatheter aortic valve replacement.

Two modern bio-prostheses inherit individual design treats to enhance their haemodynamic properties and improve their durability. The Trifecta valve (Abbott Laboratories, Abbott Park, IL, USA) is a stented bioprosthetic aortic valve designed for sutured supra-annular implantation. It features specific external leaflet mounting of a single bovine pericardial strip on a stent frame to improve the effective opening area (EOA) and to lower the trans-prosthetic gradients. The Intuity valve (Edwards Lifesciences, Irvine, CA, USA) is a supra-annular bioprosthesis, based on the stent frame and leaflet mounting of its predecessor, the Perimount Magna ease valve (Edwards Lifesciences). It incorporates a sub-annular balloon-expandable covered cloth stent and requires only 3 sutures for implantation, thus avoiding numerous pledgets in the outflow tract. Both valves present with favourable trans-prosthetic gradients and EOA after implantation in early to intermediate follow-up [2–6]. Anecdotal reports of early SVD of the Trifecta valve emerged within the last years despite the compelling initial haemodynamics. Several late studies were trying to shed light onto this topic and showed controversial results [7–11]. We aimed to analyse and compare the rates of SVD and re-interventions at long-term follow-up between patients undergoing SAVR with both low-gradient bio-prostheses in a large single-centre cohort.

METHODS

Ethical statement

Data collection and prospective follow-up (FU) for study purposes were approved by the ethical committee of the Medical University of Vienna (ethical board numbers 1180/2019 and 1861/2016; date of approval 16 April 2019 and 20 October 2016). All patients who underwent prospective clinical follow-up singed a written informed consent form after thorough explanation of study purposes.

Patients

Between April 2011 and May 2020, 1118 patients underwent successful implantation of either a Trifecta (n = 346) or an Intuity (n = 772) valve in aortic position at our institution. The Medical University of Vienna is a large tertiary referral centre for cardiac surgery. All patients, who underwent SAVR due to aortic stenosis or combined aortic disease, were included within the study. Patients operated due to sole aortic regurgitation or active endocarditis were excluded from the study population, as both indications are contraindications for implantation of an Intuity valve. Retrospective data were obtained using the institutional surgical database, which is part of the nationwide Austrian quality-assessment tool for cardiac surgery and continuously updated by a specific database administrator. Regular monitoring is performed annually, and its mortality data are in accordance with the national institute for statistics. Patients were regularly contacted postoperatively for longitudinal follow-up including echocardiographic exams.

End point definitions

The primary end point of the study was SVD as defined by the current joint STS/AATS/EACTS guidelines [12] including severe dysfunction of the operated valve as determined by necessary reintervention due to structural defects, clinical examination including echocardiography or autopsy. Echocardiographic criteria were based on a standardized definition of SVD [13] and were defined as follows: trans-prosthetic mean pressure gradient (MPG) >40 mmHg, aortic valve opening area smaller than 1 cm², maximum trans-prosthetic flow-velocity >4 m/s or presence of severe valvular regurgitation. Severe bioprosthetic dysfunction in case of endocarditis or due to non-structural defects (non-SVD) were not counted as SVD. Secondary study end points were re-interventions of the aortic valve and postoperative mortality. Clinical adverse events such as endocarditis, thromboembolic and bleeding events were defined according to mentioned guidelines [12]. Severe prosthesis–patient mismatch (PPM) was defined as an indexed EOA of <0.65 cm² calculated by dividing the EOA by the patient’s body surface area [14].

Statistical methods

Categorical variables are described by absolute numbers and percentages, continuous variables by the mean (± standard deviation) in case of normal distributions, or the median (interquartile range) in case of non-normal distributions. The chi-square test and Fisher’s exact test were used to compare categorical variables between the 2 valve type cohorts. The two-sample t-test was used for group comparisons of normally distributed continuous variables, and the Mann–Whitney U-test to compare non-normally distributed quantitative variables. The inverse Kaplan–Meier method was applied to calculate the median (interquartile range) follow-up times [15]. Cumulative incidence functions (CIF) were used to estimate the probabilities of SVD, endocarditis and non-SVD, considering death and aortic valve reintervention as competing events. In addition, the probabilities of aortic valve reintervention, late embolic event and late bleeding event were estimated by the CIF with death as a competing event. The Gray test was used to test for CIF differences between the 2 valve type cohorts. Univariable and multivariable Cox proportional cause-specific hazards regression models were used to evaluate the unadjusted and adjusted valve type effect on the primary and secondary end points SVD and aortic valve reintervention, respectively. Within these Cox regression models, a potential time-varying valve type effect was tested and included in case of statistical significance. Due to the small number of events, only 3–4 prognostic factors could have been included in the multivariable regression models to adjust for potential baseline differences of the 2 valve type cohorts. However, to prove robustness of the results achieved, a propensity score analysis was performed, considering the most relevant baseline characteristics as explanatory variables in a multivariable logistic regression model with valve type as binary outcome variable. Baseline characteristics age, sex, body mass index, active or former smoker status, hypertension, dyslipidaemia, diabetes, atrial fibrillation (AFIB), chronical obstructive lung disease, EuroSCORE II, preoperative creatinine, indication for aortic valve surgery, performance of concomitant procedures, preoperative New York Heart Association status and preoperative left ventricular ejection fraction were included in the propensity score model. The inverse probability of received treatment weighting (IPTW) method [16] was then applied to compare the 2 valve type cohorts with respect to the end points SVD and aortic valve reintervention, respectively. Again, univariable Cox proportional cause-specific hazards regression models were used. All-cause mortality was evaluated using the Kaplan–Meier method and survival curves were compared using the log-rank test. To evaluate the potential valve type effect on all-cause mortality, univariable and multivariable Cox proportional hazards regression models were performed. The previously mentioned baseline characteristics and the time-dependent variable reintervention were included in the multivariable regression model. As before, time-varying effects were tested and included in case of statistical significance. Two-sided P-values <0.05 were considered as indicating statistical significance. Analyses were explanatory in nature and there was no prespecified plan to adjust for multiple comparisons. The software SAS (SAS Institute Inc. 2016; Cary, NC, USA) was used for statistical calculations.

RESULTS

Patients and surgical data

Of 1118 patients operated between April 2011 and May 2020, 1070 patients were included (Trifecta n = 298, Intuity n = 772). Thirty-two patients were excluded due to sole aortic regurgitation and 16 patients were due to endocarditis as indication for surgery. Implanted valve models were the post-market Trifecta, the Trifecta GT, the pre-market Intuity quick-connect and the post-market Intuity Elite system. The mean age of Trifecta and Intuity recipients was 72.1 ± 8.5 years and 73.4 ± 7.6 (P = 0.019), respectively, and 61.4% in the Trifecta cohort were male compared to 54% (P = 0.029) in the Intuity cohort. Demographics and preoperative baseline characteristics are summarized in Table 1.

Considering the primary end point SVD, the median follow-up was 4.3 years (quartiles 2.1–6.3) with 1017 patient years for the Trifecta cohort and 3.2 years (1.2–5.1) with 2436 patient years for the Intuity cohort. The median follow-up with respect to overall survival was 6.1 years (5.0–7.1) for the Trifecta cohort and 4.0 years (2.2–5.8) for the Intuity cohort.

Procedural details

Concomitant procedures were performed more frequently in Trifecta patients with 64.4% (n = 192) than in Intuity patients with 43.4% (n = 335, P < 0.001). Thirty-seven patients had already underwent previous cardiac surgery before SAVR with one of the study valves (Trifecta 3.4%, n = 10; Intuity 3.5%, n = 27, P = 0.895). Based on implanted valve sizes, estimated severe PPM was significantly higher in the Intuity cohort with 58 patients and no patients in the Trifecta cohort (P < 0.001). Procedural details are summarized in Table 2.

SVD

A total of 27 cases of SVD have been observed in the overall cohort: 23 of them in the Trifecta group and 4 in the Intuity group. Competing risk analysis revealed a significantly higher cumulative incidence of SVD in the Trifecta group (P < 0.001); Incidence of SVD at 5 years was 4.49% [95% confidence interval (CI): 2.07–8.32] in the Trifecta group vs 1.04% [0.27–2.96] in the Intuity group. At 7 and 8 years, incidence of SVD increased to 23.78% [14.56–34.28] and 29.39% [18.15–41.53] in the Trifecta group and remained at 1.6% [0.5–3.98] in the Intuity group (Fig. 1A).

Valve type, logarithmic-transformed preoperative creatinine and sex emerged as independent prognostic factors of SVD in multivariable analyses (Table 3). Sensitivity analyses applying the IPTW method, based on the propensity score model, revealed comparable results with a statistically significant valve type effect [Trifecta hazard ratio (HR) 11.90, 95% CI 3.18–45.45; P = 0.0002].

Aortic valve re-interventions

In the study period, we observed 23 aortic valve related re-interventions in the Trifecta cohort and 22 re-interventions in the Intuity cohort. Cumulative incidence of re-interventions at 1 and 5 years was 1.85 (95% CI: 0.7–4.05) and 3.87 [1.87–6.99] in the Trifecta group and 2.30 [1.37–3.62] and 3.02 [1.83–4.67] in the Intuity group. At 7 and 8 years, it increased to 22.26 [12.86–33.27] and 37.28 [20.26–54.33] in the Trifecta group and to 5.65 [2.98–9.53] in the Intuity group (Fig. 1B). A significantly higher cumulative incidence of re-interventions was observed in the Trifecta cohort (P = 0.004).

Valve type emerged as the only independent prognostic factor for re-interventions with a time-varying effect (P < 0.001) and a Trifecta HR of 2.78 [95% CI 1.42–5.45] at 12 months, of 3.92 [1.83–8.40], 4.60 [1.97–10.71] and 5.32 [2.08–13.60] at 3, 5 and 8 years, respectively (Table 4). Sensitivity analyses using the IPTW approach resulted in a statistically significant valve type effect with a HR of 2.52 [95% CI 1.29–4.90, P = 0.007].

Procedural outcomes and prognostic factors for mortality

The 30-day mortality in the overall cohort was 1.8% (n = 19) with 12 patients (4.03%) in the Trifecta group and 7 patients (0.91%) in the Intuity group (P = 0.0005). Freedom from overall mortality at 1 and 5 years was 88.05% [95% CI 83.76–91.27] and 72.49% [66.80–77.37] in the Trifecta cohort and 95.05% [93.11–96.46] and 78.95% [74.82–82.48] in the Intuity cohort. At 8 years, it was 55.73% [46.58–63.92] in the Trifecta group and 58.28% [49.57–66.01] the Intuity group. Overall survival did not differ significantly between groups (log-rank test: P = 0.052) (Fig. 2). A Cox proportional hazards regression model was created to identify prognostic factors for survival including valve type, age, sex, body mass index, active or former smoker status, hypertension, dyslipidaemia, diabetes, AFIB (or flutter), chronical obstructive lung disease, EuroSCORE II, preoperative creatinine, indication for aortic valve surgery, performance of concomitant procedures, preoperative New York Heart Association status, aortic valve re-interventions and preoperative left ventricular ejection fraction. Age (HR 1.04; CI 1.02–1.06), diabetes (HR 1.49; CI 1.13–1.97), AFIB (HR 1.37; CI 1.01–1.85), log-transformed EuroSCORE II (HR 1.66; CI 1.31–2.12), log-transformed preoperative creatinine (HR 1.98; CI 1.42–2.77) and aortic valve reintervention (HR 6.99; CI 4.18–11.68) emerged as statistically significant prognostic factors for postoperative mortality. Furthermore, a statistically significant time-varying effect of the valve type resulted from the multivariable Cox regression model with a HR (Trifecta) of 1.54 [95% CI 1.00–2.39] at 3 months, 1.01 [0.74–1.39] at 1 year and 0.62 [0.43–0.91] at 5 years. Early and late clinical outcomes are summarized in Table 5.

Haemodynamic measurements

Echocardiographic assessment of trans-prosthetic MPG across all valve sizes showed a mean MPG of 10.5 ± 5.2mmHg in the Trifecta group and a MPG of 12.3 ± 5.4mmHg in the Intuity group at 1 month follow-up. Mean gradients at 1, 3 and 5 years were 13.1 ± 4.9, 16.2 ± 10.6 and 20.5 ± 17.3 in the Trifecta group and 11.1 ± 4.4, 11.4 ± 4.7 and 13.6 ± 9.1 in the Intuity group.

DISCUSSION

This study reports the first comparative analysis of SVD at long-term follow-up in patients with a Intuity and a Trifecta valve. Both bioprostheses, 1 a rapid deployment and 1 a sutured valve, are currently in wide use and are favoured for an excellent haemodynamic profile at short-to-intermediate follow-up.

Principal findings

Our study presents with the following principal findings: (i) SVD was significantly higher at long-term follow-up in patients who underwent SAVR with a Trifecta valve compared to patients with an Intuity valve; (ii) implantation of a Trifecta valve, logarithmic-transformed creatinine and female sex were positive prognostic factors of SVD in a multivariable regression model; (iii) implantation of a Trifecta valve was the only positive prognostic factor for aortic valve re-interventions; and (iv) aortic valve reintervention was a risk factor for mortality.

Bioprosthetic design has been driven to optimize the haemodynamic properties of stented valves with low postoperative gradients, based on the negative impact of elevated postoperative gradients and PPM on valve durability [17, 18]. The Trifecta valve delivers excellent early haemodynamic performance as several large-scale studies have shown in the past. In the pre-market study including 1024 patients who underwent AVR with a Trifecta, mean trans-prosthetic gradients across all valve sizes at discharge ranged from 4.1 to 9.3mmHg and remained low at 1-year follow-up with 4.7–10.7 mmHg [19]. A different collective of 824 patients implanted with the Trifecta showed average MPGs below 10 mmHg in all valve sizes but the 19-mm valve at discharge [3]. Kilic et al. [4] reported mean gradients of 6.8 ± 5.3 mmHg in the immediate postoperative period, which remained low with 10.1 ± 6.2mmHg at 1-year follow-up in their collective of 1953 trifecta patients. Despite these promising data, Biancari et al. [7] demonstrated a significantly increased risk of repeated aortic valve replacement when comparing Trifecta with Perimount (Edwards Lifesciences) recipients in the Finnish nationwide database, most of them related to SVD. Fukuhara et al. [10] showed not only a higher rate of re-interventions but also a higher rate of SVD after SAVR with the Trifecta compared to a mixed bioprosthetic collective. In the currently largest single-centre report of 2298 patients receiving a Trifecta in aortic position, comparative analysis of haemodynamic measurements with a propensity-matched cohort of Perimount (Edwards Lifesciences) recipients confirmed significantly higher rates of valve explantation at 5-year follow-up [9]. Although presenting with low gradients early on, a continuous rise of trans-prosthetic MPG was observed in the Trifecta group, while MPGs in the Perimount group were initially higher but remained relatively stable over the observed study period.

Most of these findings are in line with the results of this study; not only did we observe a significantly higher cumulative incidence of re-interventions but also of echocardiographic confirmed SVD in the Trifecta group (Fig. 3). Interestingly, a rapid increase of observed SVD events happened after 6 years, which is remarkably similar to the time pattern of observed SVDs in the cohort of Fukuhara et al. [10]. Comparison of both groups showed exceptionally low MPGs of both valve types at 1-month follow-up with a rise of the Trifecta gradients over the study period while Intuity gradients remained stable. Similar progression of gradients has been observed in comparative analysis of the Trifecta and the Perimount valve [9].

Considering the evident early failure of the Mitroflow valve (Sorin Group, Milan, Italy) with high rates of calcific degeneration, it appears that bioprosthetic design with externally leaflet mounting might be a key contributor to early bioprosthetic deterioration [20]. Experimental studies have shown that the Trifecta prosthesis demonstrates a greater widening of the struts during the ejection phase, which might result in increased leaflet stress, especially in the commissural area, where a large number of leaflet tears have been observed [21]. In addition, increased leaflet flutter of the Trifecta has been observed in an in vitro model, a phenomenon which is thought to negatively impact bioprosthetic leaflet durability [21, 22]. Others pointed out that incorrect implantation technique of the valve can attribute to early degeneration; overfitting of the valve with excessive pressure on the strut base in a too small annulus might affect the pericardial integrity in this area [23]. Given the growing number of reports of increased SVD and re-interventions rates across several Trifecta cohorts and centres worldwide, the actual influence of incorrect implantation technique on early degeneration is questionable.

The Intuity valve showed excellent data regarding freedom from SVD in our study with an observation period up to 10 years. Cumulative probability of SVD was 1.6% at 6 years and remained unchanged at 7- and 8-year follow-up with 4 cases observed in the whole study period.

Several factors may attribute to the low rate of SVD of the Intuity valve: it is the rapid deployment successor of the Perimount Magna Ease (Edwards Lifesciences) valve, with the same leaflet mounting strategy and anti-calcification treatment, which has demonstrated favourable durability in the past [24]. The Intuity presents with an even more favourable haemodynamic profile and lower gradients compared to the Perimount Magna and demonstrated favourable transvalvular gradients and a low rate of severe PPM even in smaller valve sizes [25, 26]. In a prospective study cohort of 287 patients receiving the Intuity valve, the average MPG among all valve sizes were 10.6 mmHg at discharge and remained stable at 3- and 5-year follow-up [4]. Results from our single-centre real-world collective of 700 patients showed similar results with average mean gradients of 13, 12 and 13 mmHg at 1, 3 and 5 years, respectively [5]. A specific design feature of the investigated rapid-deployment valve is the stent-based anchoring system, requiring only 3 non-pledgeted guiding sutures for positioning. The avoidance of pledgeted sutures in the subannular compartment has shown an improvement of the laminar flow and less turbulences within the prosthesis and aortic root, which could be an additional factor for reduced leaflet stress and improved durability [27]. Moreover, the radial forces of the sub-annular stent might further open and reshape the left ventricular outflow tract. The number of estimated PPM was significantly higher in the Intuity group (P < 0.001), probably attributable to 76 patients (9.84%) being implanted with a 19-mm valve compared to only 10 patients (3.36%) in the Trifecta group. Considering the higher rate of estimated PPM in the Intuity group, this highlights the results regarding the favourable durability of the rapid-deployment valve.

All-cause mortality was not significantly different between both cohorts in this study; however, aortic valve reintervention was identified as an independent predictor for overall mortality. Interestingly, a statistically significant time-varying effect of the valve type resulted from the multivariable Cox regression model, showing an initially higher HR in the Trifecta group which changed over the course of the study in favour of this group. Increased follow-up in this cohorts might provide additional information on the influence of valve type on overall mortality, as numbers of aortic re-interventions are naturally expected to increase with longer follow-up times.

Study strengths and limitations

The reported study includes 2 large single-centre cohorts of the Trifecta and the Intuity valve. Longitudinal follow-up, which included clinical visits, echocardiographic examinations and telephone follow-up in most of the patients, was performed meticulously. However, due to the retrospective design of the study, follow-up can be hindered by death of patients, relocation or patient refusal for study inclusion. As a single-centre study, our data represent a specific surgeon subset over a limited period; therefore, it might include unknown bias. This study is a comparative analysis of a sutured and a rapid deployment valve, which do inherit different design traits.

CONCLUSION

SVD was significantly more frequent in patients receiving a Trifecta valve and its implantation was an independent risk factor for the occurrence of SVD and aortic valve re-interventions. There seems to be a rapid increase of observed SVDs in the Trifecta group starting at 6 years after surgery while the Intuity valve showed exceptionally low rates of SVD throughout the study period. These results put the long-term durability of the Trifecta valve in question and need to be taken into consideration when performing bioprosthetic aortic valve replacement.

Presented at the 35th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Barcelona, Spain, 13–16 October 2021.

ACKNOWLEDGEMENT

The authors are thankful to our database administrator Ms. Manduric for her support and her continuous improvement of the institutional database.

Funding

A research grant by Edwards Lifesciences (Edwards Lifesciences, Irvine, CA, USA) was provided to conduct a long-term follow-up after Intuity valve implantation. The authors maintained full control over the study during the study period and guarantee for its design, analysis, accuracy and honesty.

Conflict of interest: Martin Andreas is a proctor/speaker/consultant (Abbott, Edwards, Medtronic) and received institutional funding (Abbott, Edwards, Medtronic, LSI). Alfred Kocher is a proctor (Edwards) and received speaker fees (Edwards). Guenther Laufer is an advisory-board member (Edwards). The other authors have no conflict of interest to declare.

Data Availability Statement

The data underlying this article will be shared on reasonable request to the corresponding author.

Author contributions

Paul Werner: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Writing—original draft. Iuliana Coti: Data curation; Resources; Writing—review & editing. Alexandra Kaider: Formal analysis; Methodology; Resources. Jasmin Gritsch: Data curation; Investigation; Project administration. Markus Mach: Conceptualization; Writing—review & editing. Alfred Kocher: Supervision; Writing—review & editing. Guenther Laufer: Supervision; Writing—review & editing. Martin Andreas: Conceptualization; Project administration; Supervision; Writing—review & editing.

Reviewer information

Reviewer information European Journal of Cardio-Thoracic Surgery thanks Tirone E. David, Michael Ibrahim, Giuseppe Santarpino and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• AFIB

  Atrial fibrillation

 
• CI

  Confidence interval

 
• CIF

  Cumulative incidence function

 
• EOA

  Effective opening area

 
• HR

  Hazard ratio

 
• IPTW

  Inverse probability of received treatment weighting

 
• MPG

  Mean pressure gradient

 
• PPM

  Prosthesis–patient mismatch

 
• SAVR

  Surgical aortic valve replacement

 
• SVD

  Structural valve deterioration