Surgical aortic valve replacement in patients aged 50–69 years—insights from the German Aortic Valve Registry (GARY)

Abstract

Open in new tabDownload slide

OBJECTIVES

The aim of this study was to analyse the outcome of patients between 50 and 69 years of age undergoing biological or mechanical aortic valve replacement.

METHODS

Data were collected from the German Aortic Valve Registry. Data were analysed regarding baseline characteristics and outcome parameters such as 5-year survival, stroke and reintervention.

RESULTS

In total,

 

3046 patients undergoing isolated surgical aortic valve replacement between 2011 and 2012 were investigated and a propensity score matching was performed. Within this period, 2239 patients received a biological prostheses, while 807 patients received a mechanical prosthesis. Mean age in the biological group was 63.07 (±5.11) and 57.34 (±4.67) in the mechanical group (standardized mean difference 1.172). In the overall cohort, there were more female patients in the biological group (32.7% vs 28.4%) and log EuroSCORE I was higher (5.41% vs 4.26%). After propensity matching (610 pairs), there was no difference in the mortality at 5-year follow-up (12.1% biological vs 9.2% mechanical P = 0.05) nor for reoperation/reintervention (2.5% biological vs 2.0% mechanical, P = 0.546). Patients undergoing mechanical aortic valve replacement suffered from a higher stroke rate 3.3% vs 1.5% (P = 0.04) at 5-year follow-up.

CONCLUSIONS

Aortic valve replacement with biological or mechanical prostheses showed similar 5-year outcomes for survival and reoperation in a propensity-matched cohort, but significantly increased stroke rate after mechanical aortic valve replacement. This could influence the choice of a mechanical valve in younger patients.

INTRODUCTION

Aortic valve disease is the most frequent heart valve disease in western countries [1]. Nowadays, following the new guidelines and the introduction of transcatheter aortic valve replacement (TAVR), surgical aortic valve replacement (SAVR) is still the gold standard in low-risk patients and non-inferior to TAVI in intermediate and high-risk patients [2]. There are still numerous patients undergoing SAVR every year, in particular young patients with bicuspid valve pathologies.

The known risk of degeneration of biological prostheses with the need for reoperation and the lack of information on long-term durability of TAVR prostheses, mechanical prostheses are still widely used in this patient group [3]. However, patients receiving mechanical prostheses carry a higher risk of thromboembolic events and lifelong anticoagulation is mandatory. As this group of patients is characterized by a long life expectancy and oral anticoagulation is difficult to manage in the long run, sustained quality of life of these patients is limited. In addition, oral anticoagulation is a risk factor for bleeding complication. Phenprocoumon/Warfarin is the only oral anticoagulation for mechanical prostheses and the intake must be controlled frequently due to its fluctuant pharmacokinetic [4]. Moreover, randomized trials have failed in attempts to employ new oral anticoagulants for patients with mechanical valves [5]. Earlier generations of biological prostheses showed a reduced durability, but various anticalcification treatment strategies have been improved, but long-term data are still lacking [6]. A relative increased use of biological prostheses is obvious [7] but the amount of evidence for using a biological valve substitute in patients younger than 65 or 60 is poor. Therefore, the aim of this study was to examine the data of the German Aortic Valve Registry (GARY) regarding survival and safety after SAVR of patients between 50 and 69 years of age.

METHODS

GARY is a prospective, collaborative, multicentre all-comers registry initiated in 2010 to monitor contemporary outcomes after treatment of aortic valve stenosis in Germany. Consecutive patients were enrolled if elective or urgent treatment (aortic valve replacement, aortic valve reconstruction and TAVR or balloon valvuloplasty) was planned and patients gave written informed consent. Detailed descriptions of GARY have been published previously [8].

Ethical statement

The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki. The Institutional Review Board of the Albert-Ludwigs-University Freiburg, Germany approved the study (EK-Freiburg: 106/10 Operations-Nr. 200192) on 5 March 2020.

Study design

Between 2011 and 2012, patients undergoing aortic valve replacement in Germany were enrolled within GARY. Exclusion criteria were patients undergoing TAVR, concomitant procedures, age <50 or >69 years of age and patients due to no valid valve type documentation (Fig. 1). Choice of the prostheses type is usually adapted to the patient's wish after information by the surgeon. In addition, we compared outcomes according to specific valve models. Within the mechanical and biological prostheses groups, the most frequently documented valve models were investigated. In the mechanical group, the ATS/Open Pivot (Medtronic, Minneapolis, MN, USA), SJM Masters (Abbott Santa Clara, CA, USA), SJM Regent (Abbott Santa Clara, CA, USA) and Carbomedics (LivaNova PLC, London, UK) were investigated, while the Perimount (Edwards Lifesciences Corp., Irvine, CA, USA), Trifecta (Abbott Santa Clara, CA, USA), Hancock (Medtronic, Minneapolis, MN, USA) and Epic (Abbott Santa Clara, CA, USA) were investigated for biological valves.

Primary outcomes were defined as the incidence of mortality, disabling stroke and reoperation/reintervention was investigated in a 5-year follow-up regarding the entire study population as well as on a propensity match. Follow-up data were obtained by contact letters or telephone interviews with the patients, their family members or data provided by registration offices. Five-year survival follow-up was complete in 95.0% (n = 2893/3046). The follow-up index (FUI) was calculated for each patient as follow-up completeness at the study end date (i.e. at 5 years) as the ratio between the actual and potential follow-up period [9]

Statistical analysis

Categorical variables are expressed as absolute and relative frequencies. Continuous variables are expressed as mean and standard deviation. Standardized mean differences were calculated to assess covariate balance. In-hospital outcomes were compared between groups in the unmatched population using Chi-squared test or Fisher’s exact test. Time to death was analysed using the Kaplan–Meier method. Cox proportional hazard models were used to investigate the relationship between the implanted valve type and time to death. The proportional hazards assumption was tested based on Schoenfeld residuals. The assumption was met for the covariate ‘valve type’. To investigate time to disabling stroke and redo/reintervention, competing risk analysis was performed with death being the competing event. Gray’s test was performed to test for differences between groups. To account for differences in baseline characteristics, the propensity score (PS) matching method was employed. The 1:1 nearest neighbour matching using the PS to receive a mechanical valve based on the covariates age, s/p stroke, atrial fibrillation, arterial hypertension and access site with exact matching on variables age, atrial fibrillation, hypertension and access site was performed. Paired tests and stratified log-rank tests were used to assess differences in outcomes between groups in the matched sample. All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk, NY: IBM Corp) and R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, Austria). The matchit package for R was used to compute the propensity match model [10] and the cmprsk package for the competing risk analysis.

RESULTS

Baseline characteristics are presented in Table 1. The distribution of the different valve types in the matched population is shown in Table 2. Out of 3046 patients, 2239 received bio-SAVR, while 807 received a mech-SAVR.

In the bio-SAVR group, 94.0% (2104/2239) had an FUI of 1. Of those, patients who died (n = 263) had a median time to death of 758 days (116.5–1270 days). In patients with an FUI < 1 (n = 135), the median follow-up time was 1105 days (389–1333 days). In the mech-SAVR group, 94.4% (762/807) had an FUI of 1. Of those, patients who died (n = 72) had a median time to death of 992 days (270–1359 days). In patients with an FUI < 1 (n = 45), the median follow-up time was 1100 days (378–1183 days).

The mean age in the bio-SAVR group was 63.07 (±5.11) and 57.34 (±4.67) in the mech-SAVR group. In the bio-SAVR group, there were more female patients (32.7% vs 28.4%). Before matching risk stratification was calculated by the log euroSCORE I and STS PROM and was significantly higher in the bio-SAVR group: log euroSCORE I 5.41% vs 4.26%, and the STS 2.00% vs 1.65%, respectively. Concomitant coronary artery disease (16.8% vs 13.1%) s/p stroke (1.2% vs 0.4%) and cerebral peripheral artery disease (5.1% vs 2.4%) was more prevalent in the bio-SAVR group. After PS matching, we had 2 groups with 610 patients each.

Survival

The estimated cumulative Kaplan–Meier 5-year mortality in the unmatched population was 12.0% in the bio-SAVR group versus 9.2% in the mech-SAVR group (P = 0.03, number of events 263/2239 vs 72/807). Estimated Kaplan–Meier mortality rates after 1 year were 4.1 vs 2.5% (P = 0.04, number of events: 91/2239 vs 20/807) and after 3 years was 8.2% vs 5.4% (P = 0.01, number of events 181/2239 vs 43/807). After 1:1 PS matching, 610 patients in each group remained and were analysed. There was no statistically significant difference regarding mortality at 5-year follow-up (12.1% bio-SAVR vs 9.2% mech-SAVR, P = 0.0502, number of events 72/610 vs 55/610) (Fig. 2). One-year mortality was 4.8% vs 2.3% (P = 0.03, number of events 29/610 vs 14/610) and 3-year mortality was 8.8% vs 5.1% (P = 0.01, number of events 53/610 vs 31/610).

There is no significant difference between the different mechanical models with respect to survival, while differences between biological valve models can be seen in the unmatched data (Fig. 3, Supplementary Material, Fig. S8). Regarding postoperative oral anticoagulation or antiplatelet therapy in biological prosthesis recipients, 58.4% were on antiplatelet medication and 57.9% on oral anticoagulation. About 95.5% were on antiplatelet therapy or oral anticoagulation and 18.9% were on antiplatelet and anticoagulation in our study population.

PS matching resulted in a matched population of n = 610 patients per group; therefore, the populations had a meaningful sample size. The PS was well-balanced between both groups (Supplementary Material, Fig. S4). The resulting groups were well-balanced regarding baseline covariate as indicated by a standardized mean difference <0.1 (Table 1). Postoperatively, there was no difference regarding classical postoperative complications such as myocardial infarction, new onset of atrial fibrillation, permanent pacemaker implantation and transfusion of red blood cells (Table 3).

Disabling stroke

Stroke rates after 1 year were 1.1% in the bio-SAVR group and 2.1% in the mech-SAVR group (P = 0.04, number of events 25/2239 vs 17/807) and after 5 years were 2.4% vs 3.5% (P = 0.08, number of events 52/2239 vs 28/807). In the matched population, stroke rates after 1 year were 0.5% vs 1.8% (P = 0.03, number of events 3/610 vs 11/610) and after 5 years were 1.5% vs 3.3% (P = 0.04, number of events 9/610 vs 20/610).

Redo surgery

In the unmatched population, the rates of redo-surgery or reintervention 1 year after the index-procedure were 1.1% vs 0.4% (P = 0.07, number of events 24/2239 vs 3/807) and after 5 years were 3.6% vs 1.9% (P = 0.03, number of events 82/2239 vs 16/807). In the matched cohort, these rates after 1 year were 1.0% vs 0.3% (P = 0.15, number of events 6/610 vs 2/807) and after 5 years were 2.5% vs 2.0% (P = 0.55, number of events 15/610 vs 13/610) In a sub-group analysis the biological valve models showed significant different redo-rates (P = 0.014) (Supplement Material, Fig. S9).

DISCUSSION

The optimal choice of prosthetic heart valves for patients between 50 and 69 years is still a debate. Therefore, the chosen prosthetic implant still represents a compromise between structural deterioration of bioprosthetic valves requiring redo intervention or mechanical devices promising lifelong durability with the downside of lifelong anticoagulation [11]. The age limit of 65 years was considered as an arbitrary point, from which onwards a bioprostheses for elderly people may have advantages as compared to the adverse events associated with mechanical prostheses [12].

In our study, data show an overall excellent 5-year survival, independent of the type of prostheses patients received. The higher rate of disabling stroke in the mechanical group is surprising—in particular in a comparatively young low-risk patient group as one would expect that adherence to anticoagulation in this patient group is acceptable. In addition, this trend persisted after PS matching and was significant at 5-year follow-up. With respect to other studies in this age group, Chiang et al. [13] with 4235 patients and 1001 propensity-matched pairs described the same findings regarding survival but no significant differences in the stroke rate. In contrast, Rodríguez-Caulo et al. [14] saw a higher rate of major bleeding in patients with a mechanical prosthesis (P = 0.004) in 15-year long-term follow-up with no significant differences regarding stroke rates whilst analysing nationwide data from Spain with 1443 patients [6]. The amount of evidence in Europe is based on Glasers study with a longer survival in patients receiving mechanical prostheses [6]. The SPAVALVE study, based on 5215 patients between 50 and 65 years, showed no difference in mortality at 15 years (74% survival), a significant higher bleeding rate (P = 0.04) and higher stroke rates for mechanical prostheses, but not significant (P = 0.07) [15]. Goldstone et al. [16] found a benefit in the long-term mortality with mechanical prostheses until 55 years of age. The only randomized trial in this field by Stassano et al. [17] showed both equal long-term survival rates as well as comparable rates of thromboembolism, bleeding, endocarditis and major adverse prostheses-related events, while patients with bioprostheses faced a significantly higher risk of valve failure and reoperation.

Regarding the risk of disabling strokes, one can speculate that patients with mechanical prostheses have an increased thromboembolic risk due to the thrombogenicity of the surface of the prostheses itself as well as a higher probability of bleeding complications as a result of mandatory anticoagulation. The achievement of proper anticoagulation intraoperatively and/or early postoperatively can be challenging as heparin, protamine and other agents that are of relevance for coagulation and clotting are frequently administered. In this setting, patients who received a mechanical valve may in fact be at higher risk. However, regarding different mechanical models, no difference was recorded in our study. Atrial fibrillation, also an independent risk factor for an increased rate of strokes, occurred similar in both groups after propensity score matching. However, strokes did not affect the mortality rate in this group of patients in our study. In Germany, it can be assumed that most patients are anticoagulated with phenprocoumon in accordance to the guidelines [18]. Although the level of anticoagulation in mech-SAVR is usually determined on an individual basis with an INR of 2.0 – 3.5, little is known about the definite INR regimens, also whether a self-assessment or an expert coagulation centre followed up these patients [19]. Some mechanical models allow a further dosage reduction of phenprocoumon towards INR 1.5–2.5 in combination with low-dose aspirin [20]. Studies on new oral anticoagulants as an alternative to phenprocoumon tried to reduce the risk for patients with mech-SAVR with more [21] or less [5] success, but none has been approved to be used. With regard to the guidelines, patients with mechanical prostheses are the only ones (alongside with patients with mitral stenosis) still dependent on phenprocoumon [18]. Since strokes can be triggered by anticoagulants, it should be emphasized that patients with atrial fibrillation and aortic valvulopathy will need—as already provided in the current guidelines [22]—to focus on the treatment of the latter to enable a dosage reduction of phenprocoumon or complete suspension in case of the use of biological implants.

Regarding a possible reoperation, after adjustment, there is no difference in the first 5 years between bio- and mech-SAVR, which one would not expect, since the valves are tested for this purpose in 200 million cycles—an equivalent of 5-year lifetime—according to standard ISO 5840-1:2021. In a longer run, studies have displayed an advantage of mechanical prostheses in young patients [23]. Reasons for a decreased use of mechanical prostheses could be a gentler redo with TAVR as valve-in-valve procedure and a restrained position regarding long-term damage by phenprocoumon.

Therefore, the risk for reoperation or reintervention needs to be taken into consideration when choosing the appropriate implant and one has to develop a long-term strategy for a younger patient receiving a biologic valve implant from the start. Our study also demonstrated a difference in risk between different biological models, in particular, a specific biological model was associated with a reoperation rate >10% at 5-year follow-up which deserves attention. Other studies show that the risk linked to reoperation lies around 5% [24] and is not negligible [25]. Due to the risk of reoperation, the second procedure often consists of a valve-in-valve procedure, so the patient can be spared a second thoracotomy [26]. However, little is known about the true effectiveness of valve-in-valve procedures on long term, such as the incidence of prostheses–patient mismatch resulting in dyspnoea and as a consequence a severe worsening of the quality of life for the respective patient [26]. In this context patient–prostheses mismatch is a central point of discussion and as a consequence of lack of long-term follow-up and increased gradients in small valves cannot be conclusively answered. Other complications are strokes due to valve-in-valve procedures [27] and limited coronary access [28].

These problems have to be solved in particular, if this is supposed to be a lifetime therapy for younger patients. The choice of valve type can also have a significant impact on the quality of life of the patient [29]. During the decision between mechanical or biological prostheses, it is necessary to consider the characteristics of the patient, the need for anticoagulation due to other causes as well as his risk factors in the event of a reoperation. An active involvement of the patient in the decision-making process leading to the prosthetic valve selection might positively affect their life perception as well as their level of compliance (shared decision-making) [30].

In conclusion, our study has shown that patients aged between 50 and 69, undergoing aortic valve replacement have a very good overall survival and an acceptable rate of reoperation comparable with contemporary results. Patients with mechanical prostheses have a higher risk of immediate, short and mid-term term stroke even at 5 years. Furthermore, a longer follow-up will allow us to respond to the risk of reoperation and long-term stroke rates in these patients.

Limitations of the study

The main limitation of this study is the follow-up period of 5 years that only allows the presentation of so-called ‘mid-term’ results. Unfortunately, no echocardiography follow-up is available in the GARY registry. Different approaches of different centres may have affected our results. Stroke, as well as redo/reintervention, was self-reported by patients or their family members. The type of stroke whether it occurred due to intracerebral bleeding or due to thromboembolic events was also not differentiated. Data on bleeding during follow-up were unfortunately not collected in GARY.

Therefore, further observation is mandatory.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Funding

The responsible body of the registry is a non-profit organization named Deutsche Aortenklappenregister gGmbH funded by the German Society for Thoracic and Cardiovascular Surgery and the German Cardiac Society. The registry receives financial support by unrestricted grants from medical device companies (Edwards Lifesciences, Medtronic, Abbott, Boston Scientific), donations from the Dr Rolf M. Schwiete Foundation and is funded by the German Center for Cardiovascular Research (DZHK) and the German Ministry of Health.

Conflict of interest: The authors declare no conflict of interest.

Data availability statement

There are no new data associated with this article.

Author contributions

Ferdinand Vogt: Conceptualization; Formal analysis; Investigation; Methodology; Project administration; Visualization; Writing—original draft; Writing—review & editing. Giuseppe Santarpino: Conceptualization; Formal analysis; Investigation; Methodology; Project administration; Supervision; Validation; Writing—original draft; Writing—review & editing. Buntaro Fujita: Data curation; Formal analysis; Software; Validation; Visualization; Writing—original draft. Christian Frerker: Formal analysis; Project administration; Supervision; Validation; Writing—original draft; Writing—review & editing. Timm Bauer: Supervision. Andreas Beckmann: Project administration; Supervision. Raffi Bekeredjian: Supervision. Sabine Bleiziffer: Investigation; Supervision. Helge Möllmann: Supervision; Writing—review & editing. Thomas Walther: Investigation; Supervision. Friedhelm Beyersdorf: Investigation; Supervision. Christian Hamm: Supervision. Andreas Böning: Investigation; Supervision; Writing—review & editing. Stephan Baldus: Investigation; Supervision; Writing—review & editing. Stephan Ensminger: Data curation; Formal analysis; Investigation; Methodology; Project administration; Supervision; Validation; Visualization; Writing—original draft; Writing—review & editing. Theodor Fischlein: Investigation; Methodology; Supervision; Writing—review & editing. Dennis Eckner: Conceptualization; Formal analysis; Investigation; Methodology; Project administration; Writing—original draft; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Pietro Giorgio Malvindi, Emiliano A. Rodriguez-Caulo and the other anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• FUI

  Follow-up index

 
• GARY

  German Aortic Valve Registry

 
• PS

  Propensity score

 
• SAVR

  Surgical aortic valve replacement

 
• TAVR

  Transcatheter aortic valve replacement

Author notes