Accuracy of the Logistic EuroSCORE in Predicting Long-Term Survival Following Isolated Aortic Valve Replacement

Objective: To assess the ability of the logistic EuroSCORE to predict long-term mortality of patients undergoing isolated Surgical Aortic Valve Replacement (SAVR). Methods: A retrospective review of all patients undergoing SAVR between September 1999, and March 2018 was done. Results: 2018 patients were eligible for inclusion in the study. Patients were grouped accord-ing to risk: low (n = 506), intermediate (n = 609), and high-risk (n = 903) depending on their logistic EuroSCORE values. The 30-day mortality of the low-risk group was 0.47%. The one-, five-, 10-, 15-, and 20-year mortality was 1.66%, 4.9%, 14.9%, 24.3%, and 43.8%, respectively. Intermediate-risk group 30-day mortality was 0.66%. The one-, five-, 10-, 15-, and 20-year mortality was 3.28%, 11.9%, 32%, 54.8%, and 82.6%, respectively. The 30-day mortality of the high-risk group was 3.99%. The one-, five-, 10-, 15-, and 20-year mortality was 8.2%, 27%, 55.4%, 78.6%, and 87%, respectively. Conclusion: Our results confirm that the lES is accurate in predicting long-term mortality outcomes of SAVR. This real-world data provides evidence of the potential usefulness of the EuroSCORE to help the heart team and patients decide on appropriate interventions for aortic stenosis.

with age [1]. Estimates are 0.2% in those aged 50 to 59, and 1.3% in the 60 to 69 age-group [1]; increasing to 9.8% in octogenarians [1]. Symptom onset is associated with poor prognosis and mortality rates approach 50% at two years if left untreated [2]. Symptoms are an important determinant in the timing of surgical management. Patients usually present with dyspnoea, chest pain, and syncope [3] [4]. Severe AS is defined as the valve area less than 1 cm 2 , mean gradient is more than 40 mmHg, and peak aortic jet velocity is more than 4 m/s. 3 Current guidelines recommend Aortic Valve Replacement (AVR) for symptomatic patients with severe AS and those who are asymptomatic with severe AS and left ventricular dysfunction [3] [4]. AVR is also recommended for patients with moderate AS undergoing other cardiac surgery [4].

The EuroSCORE
Risk stratification aids clinical decision-making and informed consenting of patients [3] [4]. Cardiac surgical risk can be predicted using the European System for Cardiac Operative Risk Evaluation (EuroSCORE, ES) [4] [5]. The first version developed in 1999, the additive EuroSCORE (aES), involving simple bedside calculations. In 2003, the logistic EuroSCORE (lES) was introduced as a more reliable tool, particularly for high-risk patients, i.e., patients with a score greater than or equal to six points [5] [6]. The low-and intermediate-risk patients score less than three and greater or equal to three, respectively [5]. The most recent and accurate form is EuroSCORE II (ES II) [7] [8]. Nonetheless, the assessment of frailty and organ dysfunction is also imperative to estimate an individual's overall surgical risk alongside the ES [4].
The aES and lES have been validated worldwide [9]. They were additionally validated for perioperative and long-term outcomes [10]. However, studies gradually reported their poor performance [11] [12]. This could be attributed to advancements in cardiac surgery techniques and post-operative care. ES II has proved to be more accurate in predicting risks [13].

SAVR or TAVR
Surgical AVR (SAVR) is the gold standard treatment in managing aortic valve disease [14]. 80% of SAVR operations were performed on low-risk patients between 2002 and 2010 [15].
The benefit of Transcatheter AVR (TAVR) in individuals with high surgical risk is established [16]. Factors favouring TAVR include lES greater than or equal to 10%, previous cardiac surgery, or the presence of porcelain aorta [3]. TAVR patients present with lower rates of major bleeding. Yet, they have a higher risk of vascular injury and paravalvular regurgitation. SAVR and TAVR have similar in-hospital mortality [17]. The two procedures have equivalent 2-year mortality in low-risk patients [18]. The low-risk group undergoing SAVR lack major systemic comorbidities, receive a bioprosthesis, with short-term mortality of 1.7%. The Placement of Aortic Transcatheter Valves (PARTNER 1) [19] trial compared standard therapy such as balloon valvuloplasty with TAVR and re-World Journal of Cardiovascular Surgery vealed a 20% reduction in mortality at one year in the TAVR group. Five-year mortality was 54.1% and 80.9% for TAVR and standard therapy, respectively. These results encouraged the PARTNER 2 [20] trial to compare TAVR and SA-VR in intermediate-risk patients. TAVR did not show superiority, nonetheless, the American College of Cardiology/American Heart Association guidelines now recommend TAVR in those with severe AS and intermediate surgical risk [4].
Subsequently, the PARTNER 3 and Evolut Low-Risk trials initiated an excitement to perform TAVR in low-risk patients [21] [22]. PARTNER 3 reported inferiority of SAVR at one year [22]. Frailty or organ system compromise would move the patient to intermediate-risk regardless of a low-risk score [4]. The limitations of unknown long-term outcomes in TAVR remain substantial [23], while there is ample data on the long-term durability of SAVR showing a trend of near-normal life expectancy post-SAVR along with a clear improvement in quality of life for most patients [24]. The presence of a bicuspid valve [25] or small aortic annulus (<21 -23 mm) [26] is identified as a factor favouring SAVR. It is also more desirable in lowrisk patients with pure Aortic Regurgitation (AR) [27].

Objective
The aim of this study is to assess the ability of the logistic EuroSCORE to predict long-term mortality of patients undergoing isolated SAVR.

Patients
All patients undergoing first-time isolated SAVR, irrespective of pathology at the Blackpool Victoria Hospital between September 1999, and March 2018 were considered for inclusion in the study. Those who lacked stated lES values were excluded. Depending on their predictive lES, patients were divided into low-risk (<3), intermediate-risk (≥3), and high-risk (≥6) groups. Patient data was collected from the independently validated NICOR (National Institute for Cardiovascular Outcomes Research) database. Our hospital database was used to access patient details retrospectively.

Statistics
Categoric variables were expressed in absolute and percentage values and analysed using chi-square test. Continuous variables were analysed for normality of distribution and values were expressed in means (±standard deviation, SD) and medians. T-test or appropriate non-parametric tests were used for comparing continuous data. Statistical significance was set at p less than 0.05. The Kaplan-Meier estimator was used for survival analysis.

Patient and Public involvement
There was no patient involvement in the design and conduct of this study.

Study Population
The starting population consisted of 2606 patients. 588 patients were excluded due to incomplete lES values. 2018 patients were eligible for inclusion in the study.
Patients were allocated to different risk-groups: low-risk (n = 506), intermediaterisk (n = 609), and high-risk (n = 903). The mean age of the study population was 68.3 (median 71), 43.3% of which were female. Table 1 outlines patient demographics of the study population. Table 2 shows the different distribution of EuroSCORE variables for the three risk groups.  Table 3. Figures 1-4 shows the Kaplan-Meier survival curves.

Mortality
We demonstrate significantly higher mortality in the high-risk population at all time points following surgery. There was no significant difference in mortality between the low-and intermediate-risk groups until between one and five years post-SAVR.

Discussion
The results of this methodologically robust real-world analysis confirm that the lES is accurate in predicting long-term mortality outcomes of SAVR.

Principal Findings
Patients in the high-risk group died significantly earlier than the low-risk and

Low-Risk Trials
The PARTNER 1 trial [19] established that TAVR is superior to medical therapy  [21] concluded that TAVR is an acceptable treatment strategy for this group of patients given that mortality rates were non-inferior to SAVR. The PARTNER 3 trial looked specifically at low-risk patients undergoing TAVR or SAVR. 30-day mortality in the TAVR group (0.4%) was less than in the SAVR group (0.9%). Results only became significant at one-year post-AVR mortality (1.1% and 2.4% for TAVR and SAVR, respectively). be drawn for the low-risk group, necessitating data from longer-term outcomes.
Of note, the TAVR-positive outcomes were demonstrated through a carefully selected group of low-risk patients with strict criteria in the PARTNER 3 trial that were followed up for no more than two years. The key question remains around the validity of these results in the real world and if these results could be applicable to all low-risk patients, outside the careful selection of these trials particularly in the context of such favourable long-term SAVR outcomes.

SAVR or TAVR
Pooled analysis of 85 retrospective cohort studies [28] reported a survival estimate of 89.7% at two years, 78.4% at five years, 57% at 10 years, 39.7% at 15 years, and 24.7% at 20 years post-SAVR. There was also a significant difference between age groups at five-year survival. Median survival was 16 years for those under 65 years of age, and 12 years and seven years in 65 to 75 and 75 to 85, respectively. This compares to our own, whole group median survival of 12.7 years. These records show that age is a strong factor in mortality post-SAVR [28].
Other studies [29] that assessed survival post-SAVR showed 80%, 55%, 32% and 21% for five, 10, 15, and 20-year survival, respectively. We report survival outcomes of 95% at one year, 83% at five years, 68% at 10 years and 58% at 20 years. Strikingly, our one-year survival (95%) is very close to the 30-day survival (95.7%) post-TAVR described by other studies [29] reiterating the non-superiority of this procedure over SAVR. Their reported outcomes for one-, and five-year There is a gap in evidence regarding outcomes of TVR in certain populations such as those with bicuspid AV, asymmetric annular calcification, and significant multivessel disease. Another component of the TAVR that needs to be explored is the durability of the heart valve. This is particularly important if younger patients with low risk undergo the procedure [15] [16] [17].
Our data raises serious concerns regarding the expansion of TAVR to all but surgically inoperable cases. We also highlight concerns regarding the expansion of findings from the PARTNER trials to the "general" aortic valve disease population. In the absence of robust, independent, long-term durability data we encourage caution with the application and expansion of TAVR and highlight the importance of a comprehensive quorate heart-team assessment of all such cases within the context of local outcome data.

Data Sharing Statement
Data available upon request.

Patient Consent
Not applicable.

Ethical Approval
Not applicable.

Conflicts of Interest
The authors declare no conflicts of interest regarding the publication of this paper.