Introduction
Surgical aortic valve replacement (SAVR) has long been the gold standard for the surgical treatment of severe aortic stenosis (AS). Over the last decade, transcatheter aortic valve replacement (TAVR) has revolutionized the treatment of severe AS. Early studies showed a clear benefit of TAVR in prohibitive and high-surgical-risk patients (Society of Thoracic Surgery-Predicted Risk of Mortality [STS-PROM] > 8%), [
1‐
3] and intermediate-risk patients (STS-PROM 4–8%) [
4,
5]. In addition, two comparative trials among low-risk patients also reported promising results [
6,
7]. Most guidelines derive their current indications from these industry-driven randomized controlled trials (RCTs) [
8,
9].
The results of RCTs have often been criticized because of the limited comparability to “real-world” practice. The surgical community has further questioned these because of the potential conflicts of interest and the perceived “suboptimal” outcomes for SAVR [
10‐
13]. However, before criticizing the RCT outcome, it is essential for us to learn the accurate outcomes of our “real-world” practice in our own country, but real-world data comparing TAVR and SAVR in Japan are scarce [
14]. As such, the purpose of the present study was to investigate the preoperative status and clinical outcomes of patients who underwent surgical treatment for severe AS in recent years and evaluate whether we could apply the RCT results in our daily practice.
Methods
Patients
Ethics approval was obtained from the appropriate ethics committee for this retrospective study. From October 2015 to July 2019, 678 patients underwent aortic valve replacement (AVR) at our institutes. Of the 678 patients, 224 underwent AVR concomitantly with aortic or mitral surgery, left ventricular restoration, left ventricular assist device implantation, or AVR for aortic insufficiency or infective endocarditis and were excluded from this study (Supplemental Fig. 1). The remaining 454 patients who underwent AVR for severe AS were then included in this study. Patients were divided into the TAVR and SAVR groups, and clinical outcomes were compared. For the mortality and morbidity analyses, patients were categorized into 3 groups according to preoperative STS-PROM score: STS-PROM ≥ 8.0 (high-risk group), 4.0 ≤ STS-PROM < 8.0 (intermediate-risk group), and STS-PROM < 4.0 (low-risk group) (Supplemental Fig. 1). Mortality and morbidity were compared between the TAVR and SAVR groups.
Selection of treatment
The selection of treatment (TAVR or SAVR) has always been performed through discussions within the heart team, consisted of cardiologists, surgeons, anesthetists, nurses, and clinical engineers. Transcatheter approach or open surgical approach was chosen considering the patient’s age, severity of condition, frailty, cognitive function, and comorbidities. All the procedures were done under general anesthesia and with the use of transesophageal echocardiography. The approach in TAVR group (transfemoral, transapical, or other alternative approaches) were chosen through discussions within the heart team considering the anatomical characteristics of the patient. SAVR was done through median sternotomy and under cardiac arrest with cardiopulmonary bypass support.
Data collection
Data were extracted from patient charts recorded in the hospital’s computer database. Follow-up started on the day of TAVR or SAVR and was censored at 30 months. Postoperative echocardiographic study was done one week after the procedure in most of the patients. In some patients in TAVR group, echocardiography was also done 1 day after the procedure to rule out complications such as cardiac tamponade. The first postoperative echocardiographic data was used for the analysis.
Statistical analysis
Continuous variables are presented as mean ± SD, and categorical variables are presented as numbers and proportions. All continuous variables were checked for normal distribution using the Shapiro–Wilk test and a normal probability plot. For univariate analyses, normally distributed variables were compared using Student’s t test, non-normally distributed variables were compared using the Mann–Whitney U test, categorical variables were compared using Chi-squared analysis or Fisher’s exact test, as appropriate. Time-to-event analyses were performed using Kaplan–Meier estimates and compared with the use of the log-rank test. The predictors of mortality were evaluated using the Cox hazard model. Factors with a value of p < 0.10 in the univariate analysis were included in the multivariable analysis.
In the low-risk group, the comparison between TAVR and SAVR was further done using propensity score matching. From the data of 66 TAVR patients and 121 SAVR patients, a multivariable logistic regression model was used to develop propensity score for the selection of TAVR with seven variables from baseline characteristics that were significantly different between TAVR and SAVR groups. The c statistics was 0.937.
All two-sided p values less than 0.05 were considered to be statistically significant.
All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria). More precisely, it is a modified version of R commander designed to add statistical functions frequently used in biostatistics [
15].
Discussion
In 2020, the Japanese Circulation Society published a revision on the guidelines for the management of valvular heart disease [
9]. In these guidelines, a clear age cutoff for the selection of TAVR or SAVR was not set. However, these offered an index of prioritization suggesting that TAVR should be utilized in patients aged ≥ 80 years, while SAVR should be used for patients aged < 75 years. The guidelines also emphasized the importance of the heart team in decision-making, including selecting a treatment approach. In our institutes, the selection of treatment (TAVR or SAVR) has always been performed through discussions within the heart team, and as a result, the mean patient age was significantly higher in the TAVR group than in the SAVR group, while the SAVR group had more comorbidities, such as coronary artery disease, peripheral vascular disease, renal dysfunction, and low left ventricular ejection fraction. Emergent/urgent operation and concomitant coronary intervention were also performed more frequently in the SAVR group than in the TAVR group. Given these differences, the most important findings in the present study were: (1) mortality in the high-risk group was as high as 27.2% at 1 year in the SAVR group, while it was significantly lower in the TAVR group (8.8% at 1 year); (2) there was no significant difference in mortality between the TAVR and SAVR groups in the intermediate-risk group; (3) mortality in the low-risk group was significantly higher in the TAVR group than in the SAVR group, and the procedure (TAVR) was a significant predictor of mortality; (4) TAVR was less invasive compared to SAVR, as seen in the decrease in bleeding, blood transfusion requirement, new-onset atrial fibrillation, and intensive care unit stay; (5) and the performance of transcatheter valves was better compared to surgical valves, with lower peak velocity and pressure gradient, higher effective orifice area index, and less frequent PPM.
The results of mortality and morbidity analyses in the high-risk group in the present study were compatible with those of industry-driven RCTs since TAVR was comparable with, or even superior to, SAVR [
1‐
3]. In the PARTNER trial, which compared TAVR and SAVR in high-risk patients [
2], the 1-year mortality of the SAVR arm was 26.8%, similar to that of the SAVR group in our study (27.2%). Furthermore, in the said trial, the patients’ age and STS-PROM scores in the TAVR group were 83.6 ± 6.8 years and 11.8 ± 3.3, respectively, while the 1-year mortality was 24.2% [
2]. In contrast, the 1-year mortality of the high-risk TAVR group in the present study was much lower (8.8%), although the age (86.5 ± 4.9 years) and STS-PROM (12.9 ± 5.2) were higher.
In 2016 and 2017, the results of two industry-driven RCTs comparing TAVR and SAVR in intermediate-risk AS patients were published [
4,
5]. Both studies revealed the noninferiority of TAVR to SAVR. The intermediate-risk group’s mortality and morbidity analysis results in the present study were also compatible with these two RCTs. Therefore, it is indisputable that TAVR should be the first choice of treatment for severe AS in high- and intermediate-risk patients.
On the other hand, the mortality and morbidity analysis results in the low-risk group in the present study slightly varied from those of industry-driven RCTs [
6,
7]. In the present study, postoperative mortality was significantly lower in the SAVR group than in the TAVR group. Even though the patients were older and the frequency of cerebral vascular disease/carotid disease and preoperative STS-PROM were higher in the TAVR group, these preoperative factors were not detected as significant risk factors of mortality in the multivariable Cox hazard model analysis. Only the preoperative factor of “procedure = TAVR” was detected as a significant risk factor for mortality. The result was the same even after the propensity score matching. As for the analysis of postoperative factors, mild or greater paravalvular leakage was detected as a significant risk factor for mortality. Over 80% of the TAVR patients had trivial or greater paravalvular leakage, with 18.2% having mild or greater paravalvular leakage, possibly explaining why TAVR was a risk factor for mortality. It is widely known that moderate and severe aortic regurgitation is associated with increased mortality [
4,
16].
In the PARTNER 3 trial, the 1-year mortality of the SAVR group, which had a mean age of 73.6 ± 6.1 years and a mean STS-PROM of 1.9 ± 0.6, was 2.5% [
6]. In contrast, in the low-risk SAVR group in the present study, which had a mean age of 71.6 ± 7.8 years and a mean STS-PROM of 2.2 ± 1.0, the 1- and 2-year mortalities were 0% and 0.9%, respectively. In our “real-world” experiences, patients under 75 years without any comorbidities that increase the STS-PROM score seldom die early after aortic valve replacement surgery. With a 1-year mortality of 2.5%, the PARTNER 3 trial could not dispel the impression that the outcomes in the SAVR arm were “sub-optimal” [
13]. Furthermore, in the said trial, 5 patients died within 30 days, with 3 dying due to “pulseless electrical activity arrest,” 1 due to respiratory failure, and 1 due to sepsis. One patient also died due to “failure to wean from extracorporeal membrane oxygenation” on day 31 [
6]. One can only wonder what happened to these six patients intraoperatively.
Study limitations
The present study had several limitations, including its retrospective nature. As stated above, the indications for TAVR or SAVR were determined through discussion within the heart team and the preoperative patient characteristics were significantly different between the TAVR and SAVR groups. In this study, risk stratification was done only by the STS-PROM, although patients underwent each procedure due to multiple reasons. For example, some patients underwent SAVR despite high STS-PROM because of hemodialysis, complex coronary artery disease, etc., while other patients underwent TAVR despite low STS-PROM because of frailty and so on. Therefore, it was impossible to compare the results of the present study and previously published industry-driven RCTs. Similar to the industry-driven RCTs, the most important data that the present study failed to show was the patients’ long-term outcomes, including data on prosthetic valve deterioration. The data would be essential in deciding whether the indication for TAVR should be expanded to young and low-risk patients. Hopefully, future studies will help clarify this.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.