Outcomes of surgical aortic valve replacement over three decades

,

In a large SAVR cohort, relative survival is close to 90% at 10 years.This excellent long-term result reinforces the role of SAVR, especially in younger lowrisk patients with long life expectancy.

PERSPECTIVE
These excellent long-term results, especially in the younger low-risk patient population with long life expectancy and lower operative risk, reinforce the role of SAVR in the treatment of aortic valve disease and serve as a benchmark for future dedicated long-term TAVR studies.
See Commentaries on pages 1752 and 1753.
Invasive treatment of aortic valve disease has been continuously evolving since the first surgical aortic valve replacement (SAVR) was performed in the 1960s. 1 Technical and procedural refinements, continuous prosthesis development, and periprocedural care improvement resulted in a substantial improvement of SAVR outcomes over the last decades. 2Concurrently, patient characteristics have changed considerably, and the comorbidity burden is increasing. 2,3he latest revolution in treating aortic valve replacement was the introduction of transcatheter aortic valve replacement (TAVR) in the early 2000s. 4Attractive for its less invasiveness, TAVR quickly became an established treatment modality for patients with aortic stenosis (AS) having high or intermediate surgical risk. 5,6More recently, clinical trial results have even challenged the role of SAVR in low-risk patients with AS. 7,8 These results forecast a new era in treating aortic valvular pathology, when optimal treatment allocation will become increasingly important.
Detailed analysis of patient and procedural characteristics, especially long-term survival after SAVR, is inevitable for informed treatment decisions.This study aimed to assess the trends in patient and procedural characteristics and the long-term survival in SAVR in a high-volume tertiary center over the last 3 decades.

MATERIALS AND METHODS Study Design and Data Collection
Adult patients undergoing SAVR between 1987 and 2016 at the Erasmus Medical Center, Rotterdam, The Netherlands, were analyzed.Patients receiving bioprosthetic or mechanical aortic valve prosthesis with or without concomitant cardiac procedures were included.Patients aged less than 18 years and patients receiving valved conduits were excluded.Baseline and procedural characteristics were collected retrospectively from electronic medical records.Survival status was obtained through the Dutch National Death Registry.
This study was conducted according to the privacy policy of the Erasmus Medical Center and regulations for the appropriate use of data in patient-oriented research, which are based on international regulations, including the Declaration of Helsinki (Institutional MEC Number: MEC-2019-0721), and patient informed consent was waived.All the authors vouch for the validity of the data and adherence to the protocol.

End Points and Definitions
The primary end point was the differences in baseline and procedural characteristics in the overall and primary isolated SAVR cohort, in three 10-year time periods according to the date of SAVR (period A: 1987-1996; B: 1997-2006; C: 2007-2016).The survival in the overall and primary isolated SAVR cohort was analyzed and compared with the survival of the matched general population (relative survival).SAVR within 24 hours of establishing the indication was classified as urgent.SAVR after 24 hours was classified as (semi-) elective.Left ventricular function was classified as normal if the left ventricular ejection fraction (LVEF) was greater than 50%, as reduced if the LVEF was 30% to 50%, and as severely reduced if the LVEF was less than 30%, as measured or estimated by a trained echocardiographer.Low-, intermediate-, and high-risk patients are defined as logistic European System for Cardiac Operative Risk Evaluation of 10 or less, 10 to 20, and 20 or greater, respectively.1987-1996; period B: 1997-2006; period C: 2007-2016).Trend analysis was performed with the chi-square test for trend.

Statistical Analysis
The relative survival can be used as an estimate of cause-specific mortality.It is defined as the ratio between the observed survival and the expected survival in the general population. 9The Human Mortality Database was used to obtain the age-, sex-, and year-matched expected survival data of the general population of The Netherlands. 10 The Human Mortality Database is continuously updated and includes mortality data from the Netherlands up until 2016.Relative survival is estimated through the Ederer II method. 11,12Data management and statistical analyses were performed using SPSS 25.0 (SPSS Inc, Chicago, Ill) and R software, version 3.5 (R Foundation, Vienna, Austria).

Baseline Characteristics
Between  E1 and E2 for the isolated SAVR and the SAVR with concomitant CABG cohort.

Changes in Patient Profile Over Three Decades
wDuring the 30-year observation period, the annual number of patients undergoing SAVR per period increased, from an annual average of 91 in period A to 187 in period C (Figure 1).The mean age increased from 63.9 AE 11.2 years in period A to 66.2 AE 12.3 years in period C (P<.001).The proportion of patients aged 70 years or more increased from 35.2% in period A to 46.7% in period C (P<.001).Between periods A and C, the prevalence of diabetes mellitus in the study population increased from 7.6% to 20.5% (P < .001),hypercholesterolemia from 5.2% to 25.0% (P<.001), and chronic obstructive pulmonary disease from 7.9% to 12.1% (P <.001).The percentage of patients with previous cardiac operations (P <.001) and redo SAVR decreased (P ¼ .023).Further changes in baseline characteristics are shown in Table 1 for the overall cohort and in Tables E1 and E2 for the primary isolated SAVR and the primary SAVR with concomitant CABG cohort.From period A to period C, the incidence of concomitant tricuspid and aortic procedures increased.The proportion of patients receiving bioprosthetic valves increased significantly, from 18.8% in period A to 67.1% in period C (P <.001, Figure 2).Detailed trends regarding changes in procedural characteristics and concomitant procedures are provided in Table 2.

Trends in 30-Day Mortality and Long-Term Survival
The 30-day mortality in the overall cohort decreased from 2.7% in period A to 1.8% in period C (P ¼ .003).The 30-day mortality across 3 decades decreased, nonsignificantly, from 1.9% to 0.9% (P ¼ .190)for primary isolated SAVR, and from 4.1% to 3.0% (P ¼ .384)for primary SAVR with CABG (Table E3).The 10-year survival was 59.8% in the overall cohort, 65.5% in the isolated SAVR cohort, and 51.1% in the SAVR with concomitant CABG group (Table 3).
From period A to C, 10-year survival did not change in the overall cohort and patients receiving isolated SAVR from 62.8% to 60.3% (P ¼ .051)and 66.9% to 67.2%, respectively (Table 3).Further trends in 10-year survival in various subgroups are displayed in Table 3 and Figures E1 to E3.Further trends in survival are shown in Tables E4 and E5.

DISCUSSION
In this study, although the age, frequency of comorbid conditions, and complexity of patients undergoing SAVR increased over a 30-year period, the trends in 10-year survival remained stable or improved.Relative survival after SAVR was 85.8% (CI, 83.5-88.1)at 10 years.In patients undergoing primary isolated SAVR, the relative survival was 92.4% (CI, 89.4-95.6)and 73.8% (CI, 67.1-81.1)at 10 and 20 years, respectively.These excellent long-term results reinforce the role of SAVR in the treatment of aortic valve disease, especially in the younger low-risk patient population with long life expectancy and lower operative risk.
In our cohort, we saw a continuous increase in the number of patients undergoing SAVR.This increase is parallel to the growing number of SAVRs performed annually in Europe and the United States over the last decades, 13 and is most likely a result of a combination of factors.The ageing of the population led to an increase in the prevalence of AS in the Western countries, 14,15 and improvements in imaging might have led to an increase of patients being referred for SAVR. 16Simultaneously, expanding indications for SAVR and practice-related changes had a positive effect on the number of SAVRs performed. 5,17Of note, this trend might be halted by the growing use of TAVR in elderly patients, which can eventually lead to a decrease in the annual number of SAVRs, a recent trend already observed in some countries. 18,19he increasing frequency of comorbidities in our patient population is in accordance with the previously described changes in the profile of patients undergoing cardiac surgery. 20The prevalence of diabetes mellitus, hypercholesterolemia, and hypertension has at least doubled during the 30year observation period.Diabetes is associated with worse outcomes in patients undergoing cardiac surgery. 21Further, 31.1% of the patients in this study underwent concomitant CABG.Hypercholesterolemia and hypertension are well known to be associated with coronary artery disease.Coronary artery disease is present in up to 40% of the patients with AS undergoing SAVR and in up to 50% in SAVR

ADULT
patients aged 70 years or more. 22,23Patients with concomitant CABG reflect a population with more advanced heart disease and diminished life expectancy due to higher shortand long-term mortality compared with those undergoing isolated SAVR. 245][26] These patients should be carefully selected and directed to high-volume centers. 25

ADULT
Prosthesis choice is an important element of treatment decisions in aortic valve disease.Both mechanical and bioprosthetic valves are associated with inherent risks. 27Mechanical valves require lifelong anticoagulation associated with bleeding events, and bioprosthetic valves are prone to degeneration, necessitating a second intervention in the long term. 28In our study, a 4-fold increase in bioprosthetic valve use was observed over the last 3 decades, mimicking a worldwide trend. 28The shift from mechanical to bioprosthetic valves was most prominent in patients aged 60 to 70 years. 29Additionally, the age profile of SAVR patients changed considerably, with an increasing number of elderly patients undergoing SAVR.These patients form the bulk of the contemporary SAVR population and received almost exclusively a bioprosthetic valve.Although the first randomized controlled trial comparing bioprosthetic and mechanical valves showed better survival in patients receiving mechanical valves, 30 recent literature supports the benefit of bioprosthetic valves compared with mechanical valves in patients aged 60 years and older. 28,31lthough younger patients might benefit from bioprosthetic valves, caution is warranted. 32Valve-in-valve TAVR in prospect might be an option when considering bioprosthetic valves in younger patients. 33,34espite the increasing patient age and complexity, the 30-day mortality decreased or remained stable over the 30-year observation period in the different cohorts.This may reflect advances in surgical technique and perioperative care over the last decades. 35Although long-term actual survival after SAVR is influenced by the competing risk of mortality due to other factors, relative survival provides a good estimate of the disease-and intervention-related risks, because it compares the survival of the investigated population with the survival of the matched general population. 36laser and colleagues 37 reported a relative survival of 97% and 88% at 5 and 10 years after SAVR, respectively, and Kvidal and colleagues 23 described a 74.9% relative survival at 15 years in a large SAVR cohort.In our study, the relative survival after isolated SAVR was similar to that of the age-, sex-, and year-matched Dutch population at 5 years, greater than 90% at 10 years, indicating an excellent long-term result.However, the decrease afterward in relative survival is not negligible and emphasizes the impact of disease-and intervention-related hazards in the extended long term. 37

ADULT
The growing use of TAVR challenges the traditional role of SAVR in the treatment of aortic valve stenosis.In the light of recent trial results, the elderly SAVR population might have overlapping indications for both TAVR and SAVR in the future. 7,8In the current 5-year data regarding intermediate-risk patients with severe symptomatic aortic stenosis, there was no difference between the incidence of the composite end point of mortality and disabling stroke in patients receiving TAVR and SAVR, 47.9% and 43.4%, respectively. 38The added value even translated to the lowrisk population.Patients classified as low risk had noninferior outcomes regarding the composite end point of mortality and disabling stroke at 2 years of follow-up, 5.3% and 6.7% in TAVR and SAVR, respectively. 8Further research regarding the long-term durability of TAVR and the use of TAVR in specific patient groups, such as patients with high anatomic risk, including bicuspid morphology, dilated aortic root, heavy annular calcification, and expected future coronary access, remain warranted.Regular formal heart team discussions are recommended by the clinical guidelines. 5,6hese meetings allow for informed decisions in a multidisciplinary setting, where the preferred intervention can be discussed on the basis of the individual patient profile, local resources and expertise, and the evidence available on procedure-related risks and long-term results. 39

Study Limitations
The results presented are based on data from a single center in The Netherlands.As with all retrospective studies, inherent shortcomings related to data capture are present.In addition, our study evaluated only survival as a longterm clinical outcome, because other important clinical outcomes (eg, quality of life, structural valve dysfunction or valve-related thromboembolic, and bleeding events) were not captured in our database.The amount of patients with newer-generation valves such as sutureless valves is low, which might yield different outcomes.Other potential limitations include selective outcome reporting.

CONCLUSIONS
The present study demonstrates the patient-related changes over time in patients receiving SAVR and the excellent SAVR-related outcomes over the last 3 decades.Isolated SAVR has proven itself with excellent long-term relative survival (73.8% at 20 years in our study).The existing SAVR cohort overlaps with the expected future TAVR cohort; therefore, our findings may serve as a benchmark for future TAVR population studies.

Long-term survival in a large cohort undergoing surgical aortic valve replacement in our center during the last 30 years
We identify that the relative survival is 60% at 20-years of follow-up.These excellent long-term results reinforce the role of surgical aortic valve replacement, especially in younger low-risk patients with long life expectancy.

patients
More than 20-years of follow-up

FIGURE 1 .
FIGURE 1. Age at operation and annual number of patients undergoing SAVR over 30 years.Over 30 years, the percentage of elderly patients and the annual number of patients undergoing SAVR increased considerably.Results are reported according to the time of SAVR (period A: 1987-1996; B: 1997-2006; C: 2007-2016).A, Annual average of patients undergoing SAVR, according to the type of surgery.Y-axis represents the absolute number of patients.B, Age distribution of patients at the time of SAVR.SAVR, Surgical aortic valve replacement; CABG, coronary artery bypass grafting.
Categorical variables are presented as numbers, percentages, or proportions and compared with the chi-square test or the Fisher exact test, where appropriate.Continuous variables are presented as means AE standard deviation or median with the interquartile range and compared with the 2sample t test or Wilcoxon rank-sum test where appropriate.Patients were classified into 10-year time periods based on surgery date (period A: 1987 and 2016, a total of 4404 patients underwent SAVR with a biological (n ¼ 2301) or mechanical (n ¼ 2103) valve prosthesis.No patients were lost to follow-up for survival, with a mean follow-up of 13.8 years.Mean age was 65.5 AE 12.1 years, and 38.2% (n ¼ 1683) were female.A total of 46.3% (n ¼ 2041) required concomitant procedures, and 5.6% (n ¼ 247) had redo SAVR.The indication for operation was AS or combined AS and aortic regurgitation in most cases (83.9%).The most common comorbidities included hypertension (35.1%, n ¼ 1545), atrial fibrillation (17.6%, n ¼ 775), and diabetes mellitus (14.9%, n ¼ 656).The median logistic European System for Cardiac Operative Risk Evaluation (available since 2003; n ¼ 2605) was 5.0%, with 18.8% (n ¼ 480) of the patients having a logistic European System for Cardiac Operative Risk Evaluation of 10% or greater and 6.0% (n ¼ 153) having a logistic European System for Cardiac Operative Risk Evaluation of 20% or greater.Further baseline characteristics are shown in Table 1 for the overall cohort and in Tables

TABLE 1 .
Baseline characteristics over three decades in the overall cohort 001 The Journal of Thoracic and Cardiovascular Surgery c December 2022 Adult: Aortic Valve C ¸elik et al ADULT concomitant CABG were older compared with patients not requiring revascularization (70.1 AE 8.3 vs 65.0 AE 12.0; P < .001). 5),

TABLE 2 .
Procedural characteristics over three decades in the overall cohort Values are presented as percentages.CABG, Coronary artery bypass grafting; VD, vessel disease; MV, mitral valve; TV, tricuspid valve.The Journal of Thoracic and Cardiovascular Surgery c December 2022 FIGURE 2. Mechanical and bioprosthetic valve use across 3 decades.Absolute number of bioprosthetic and mechanical valves implanted according to patient age and time of SAVR (period A: 1987-1996; B: 1997-2006; C: 2007-2016).Note the considerable increase in patients receiving bioprosthetic valves from period A to C and the decrease in mechanical valve use above the age of 65 years.The X-axis represents the age at SAVR.The Journal of Thoracic and Cardiovascular Surgery c Volume 164, Number 6 1747 C ¸elik et al Adult: Aortic Valve

TABLE 3 .
Ten-year survival after primary surgical aortic valve replacement over three decades Values are presented as percentages.SAVR, Surgical aortic valve replacement; CABG, coronary artery bypass grafting; MV, mitral valve; LES, logistic European System for Cardiac Operative Risk Evaluation; N/A, not available.FIGURE 3. Long-term survival after SAVR.Actual survival of patients in the overall SAVR cohort (red line) and relative survival compared with the age-, gender-, and year-matched Dutch population (blue line).The relative survival after SAVR is approximately 85% at 10 and 60% at 20 years when compared with that of the matched general population.The Journal of Thoracic and Cardiovascular Surgery c December 2022 Adult: Aortic Valve C ¸elik et al Long-term survival after primary isolated SAVR.Actual survival (red line) and relative survival compared with the age-, gender-, and yearmatched population (blue line).Note the relative survival of 73.8% after primary isolated SAVR at 20 years.
FIGURE 5. Long-term actual and relative survival after primary SAVR with concomitant CABG.Actual survival (red line) and relative survival compared with the age-, gender-, and year-matched population (blue line).Note the relative survival of 41.6% after SAVR with concomitant CABG at 20 years.The Journal of Thoracic and Cardiovascular Surgery c Volume 164, Number 6 1749 C ¸elik et al Adult: Aortic Valve Long-term actual and relative survivals in the overall cohort.Long-term survival after SAVR.Actual survival of patients in the overall SAVR cohort (red line) and relative survival compared with the age-, gender-, and year-matched Dutch population (blue line).Note the relative survival of 85.8% at 10 and 60.4% at 20 years, respectively.The Journal of Thoracic and Cardiovascular Surgery c December 2022 Long-term survival after SAVR in the overall cohort according to period operated.Actual survival of patients in the overall SAVR cohort.Patients operated between 1987 and 1996 (period A) are shown with the red line; patients operated between 1997 and 2006 (period B) are shown with the blue line; and patients operated between 2007 and 2017 (period C) are shown with the orange line.Comparison within periods is done for 10 years of followup and shown as P value.Long-term actual after primary SAVR with concomitant CABG according to period operated.Actual survival of patients with primary SAVR and concomitant CABG.Patients operated between 1987 and 1996 (period A) are shown with the red line; patients operated between 1997 and 2006 (period B) are shown with the blue line; and patients operated between 2007 and 2017 (period C) are shown with the orange line.Comparison within periods is done for 10 years of follow-up and shown as P value.
FIGURE E2.Long-term survival after primary isolated SAVR according to period operated.Actual survival of patients with primary isolated SAVR.Patients operated between 1987 and 1996 (period A) are shown with the red line; patients operated between 1997 and 2006 (period B) are shown with the blue line; and patients operated between 2007 and 2017 (period C) are shown with the orange line.Comparison within periods is done for 10 years of follow-up and shown as P value.1751.e1The Journal of Thoracic and Cardiovascular Surgery c December 2022 The Journal of Thoracic and Cardiovascular Surgery c Volume 164, Number

TABLE E1 .
Baseline and procedural characteristics over 3 decades in patients undergoing primary isolated surgical aortic valve replacement Values are presented as n (%) or as mean AE SD or median (interquartile range) if otherwise stated.SD, Standard deviation; AS, aortic stenosis; AR, aortic regurgitation; euro-SCORE, European System for Cardiac Operative Risk Evaluation; IQR, interquartile range; N/A, not available; PCI, percutaneous coronary intervention; COPD, chronic obstructive pulmonary disease; LVEF, left ventricular ejection function.

TABLE E2 .
Baseline and procedural characteristics over three decades in patients undergoing isolated surgical aortic valve replacement þ coronary artery bypass grafting Standard deviation; AS, aortic stenosis; AR, aortic regurgitation; euroSCORE, European System for Cardiac Operative Risk Evaluation; IQR, interquartile range; N/A, not available; PCI, percutaneous coronary intervention; COPD, chronic obstructive pulmonary disease; LVEF, left ventricular ejection function.
1751.e5The Journal of Thoracic and Cardiovascular Surgery c December 2022

TABLE E3 .
Thirty-day mortality after primary surgical aortic valve replacement over 3 decades Values are given in percentages with (number of patients).SAVR, Surgical aortic valve replacement; CABG, coronary artery bypass grafting; MV, mitral valve; LES, logistic European System for Cardiac Operative Risk Evaluation; N/A, not available.The Journal of Thoracic and Cardiovascular Surgery c Volume 164, Number

TABLE E4 .
1-y survival after primary surgical aortic valve replacement over 3 decades Surgical aortic valve replacement; CABG, coronary artery bypass grafting; MV, mitral valve; LES, logistic European System for Cardiac Operative Risk Evaluation; N/A, not availableCABG, Coronary artery bypass grafting; LES, Logistic European System for Cardiac Operative Risk Evaluation; MV, mitral valve; N/A, not available; SAVR, surgical aortic valve replacement.1751.e7The Journal of Thoracic and Cardiovascular Surgery c December 2022

TABLE E5 .
Five-year survival after primary surgical aortic valve replacement over 3 decades SAVR, Surgical aortic valve replacement; CABG, coronary artery bypass grafting; MV, mitral valve; LES, logistic European System for Cardiac Operative Risk Evaluation; N/A, not available.The Journal of Thoracic and Cardiovascular Surgery c Volume 164, Number