Study design

This was a prospective, observational study of all consecutive patients with severe, symptomatic AS treated with elective TAVI via transfemoral access with SAPIEN 3 at the regional Hospital Universitario Virgen de la Arrixaca, a tertiary hospital and a regional referral center for cardiothoracic surgery and interventional cardiology located in Murcia, Spain. Patients received TAVI regardless of the study as part of the routine clinical practice. The recruitment stage was during the calendar year of 2018. In this study we present the observed results from the systematic djustent conducted 1 year after TAVI.

According to the ESC/EACTS guidelines,⁶ implantation decision was made by the heart team and all procedures followed the recommendations established by the manufacturer's SAPIEN 3 valve instructions for use. All patients were followed for, at least, 12 months after TAVI and systematically assessed according to the hospital clinical protocol. Written patient information was provided to each participant, and the patient's consent on data collection was signed before being included in the study that was conducted in full compliance with the recommendations guiding biomedical research in human subjects adopted by the 18^th World Medical Assembly, Helsinki, Finland back in 1964. The study protocol was approved by the assigned ethics Committee (Murcia, SARU Study; code: 2017-01. Effective date, 02/06/2017).

Clinical assessment was conducted at baseline (preoperative), post-intervention (perioperative), and 30 days, 6 months, and 1 year after the procedure. The main objective clinical variables included echocardiographic measurements (eg, paravalvular and total aortic regurgitation, left ventricular ejection fraction, mean and maximum aortic valve gradient, effective orifice area), and major clinical events automatically available in the medical records (eg, all-cause and cardiovascular mortality, stroke, bleeding complications, myocardial infarction, new-onset atrial fibrillation, major vascular complications, permanent pacemaker implantation, rehospitalization and acute kidney injury). In addition, the patients' New York Heart Association (NYHA) functional class IV was systematically registered. The patients' risk profile was characterized based on the STS risk score which was validated for the in-hospital and 30-day mortality rates following surgical aortic valve replacement.¹³ Additionally, postoperative complications were defined based on a modified version of the Valve Academic Research Consortium criteria,¹⁴ and this score was routinely applied based on the clinical protocols of our referral hospital.

Finally, length of stay (LOS)–at the hospital and the intensive care unit (ICU) settings–associated with the TAVI procedure was automatically registered for each patient based on hospital medical records and described as a secondary endpoint in this research.

Measurement of patient's health-related quality of life

A comprehensive assessment was implemented by combining a patient-reported disease-specific tool that has a higher ability to capture changes in the patient's health status during the observation period, and 2 generic tools to establish comparisons with findings from other procedures or diseases and with the Spanish normal population. Patients' health-related quality of life was evaluated at baseline and during per protocol medical visits for health management (6 months and 1 year after TAVI).

Statistical approach

First, an exploratory analysis was performed to characterise the analytic cohort presented in this manuscript. Descriptive statistics were used for continuous variables (eg, mean, standard error of measurement) and frequency tables or proportions for discrete variables. McNemar's test for dependent samples was used to compare NYHA health states at follow-up. Regarding the objective of this study, multivariate models (linear general models with repeated measures at different time points baseline, 6M, and 12M as intra-subject factors) were computed to better assess the potential benefits of both regarding clinical endpoints and the patients' HRQoL at follow-up (baseline, 6M, and 12M) while considering the patients' comorbidities and risk profile at baseline. To this end, the STS predicted risk of mortality score was included because it is a weighted index of the patients' risk robustly estimated using a Bayesian hierarchical model for both mortality and major complication events. This model considers 24 meaningful preoperative variables like age, sex, body surface area, atrial fibrillation, chronic heart failure, NYHA functional classification, chronic obstructive pulmonary disease, diabetes mellitus, need for insulin use, arterial hypertension, previous cardiac surgeries, concomitant mitral stenosis, unstable angina, previous percutaneous coronary intervention, and other variables. Based on the estimated STS score, patients were classified into 3 risk groups: high (≥ 8%), intermediate (≥ 4%), and low mortality risk (< 4%).¹³ Importantly, this score was also considered in the secondary endpoint associated with the description of the LOS (at both the hospital and ICU settings) related to TAVI procedures.

Regarding the size of the sample required to conduct the adjusted analyses described above, the estimated minimal sample size was set at 60 TAVI patients to compare within and between subject differences at 3 different time points of evaluation, effect size (f) was 0.25, statistical power (1-β), 0.9, and risk of type-I error (1-α), 0.95 assuming a weak correlation among repeated measures (0.3).

The software statistical package SPSS 27.0 for Windows (IBM Corp., United States) and the R software (The R Project for Statistical Computing, Institute for Statistics and Mathematics, Austria) were used for analysis.

Clinical outcomes

Successful implantation was achieved in 74 cases (97.37%), and 2 patients (2.63%) died within the first 30 days after the procedure (both cardiovascular causes). In this period, the observed rates of major complications or rehospitalizations were low (table 2), and permanent pacemaker implantation was required for 5 patients. One year after TAVI, no additional cardiovascular deaths were reported. However, 7 patients died of other causes (table 2). We should mention that the sample mortality rate was similar to that of a comparable general population (figure 1 of the supplementary data). According to echocardiographic measurements, significant benefits in mean and maximum gradients, and aortic regurgitation were achieved and maintained at follow-up (figure 2 of the supplementary data). Notably, among survivors, 78.4% of patients had a NYHA functional class I-II at 1 month, a benefit that was maintained at 1 year with 77.3% of patients in these functional levels.

Figure 1. New York Heart Association (NYHA) functional class at follow-up. 

Figure 2. Changes in patient's health-related quality of life at follow-up (n = 55, out of 67 survivors at 1 year; 82.1% of the sample). No differences in mean/median baseline values were seen in any of the health-related quality of life (HRQoL) measures taken between the patients who completed all the measurements and those with missing values at study period. EQ-5D-5 VAS, EQ visual analogue scale; KCCQ, Kansas City Cardiomyopathy Questionnaire; KCCQ CSS, KCCQ clinical summary score; KCCQ OSS, KCCQ overall summary score; SF-36 MCS, Medical Outcomes Study Short Form-36 Mental Component Summary; SF-36 PCS, SF-36 Physical Component Summary. 

HRQoL assessment

The patients' HRQoL across the observational period according to STS risk score at baseline is shown on figure 1. In both summary scores of the KCCQ (OSS and CSS), statistically and clinically significant improvements after TAVI were seen. In OSS cases, the mean differences reported between baseline and 6 months ranged between 18.94 points (low risk) and 29.97 points (intermediate risk) indicative of a meaningful benefit (all P < .001). Similarly, the mean differences regarding the CSS ranged between 13.03 points and 27.3 points (low and intermediate, respectively). In addition, in both scales the observed benefit was maintained at follow-up with no differences being reported at 6 months and 1 year (P > .8 in both summary scales). Remarkably, these improvements were repeated across the 3 risk groups (figure 2).

Regarding generic questionnaires, a positive impact was also seen in EQ-5D-5L VAS scores and the SF-36 Physical Component Summary, postoperatively, in all patients (P < .006 and P < .004, respectively). However, the size of detected differences was smaller considering the respective scales. No statistically significant differences were seen in the mental component summary (P = .395). In addition, values in all groups were comparable to normative population based on age and sex since baseline (mean 95% confidence interval [95%CI] at baseline: 47.44, 44.70, and 50.18; 6 months: 49.77, 47.36, and 52.17; and 1 year: 48.64, 46.04, and 50.83-; P = .52). Finally, EQ-5D-5L index values were fairly similar across all time points with a slight increase observed at 6 months (P = .054 compared to baseline) and a mild drop at 1 year, not reaching statistical significance compared to baseline values (mean 95%CI at baseline: 0.77, 0.71, and 0.83; 6 months: 0.80, 0.75, and 0.85; and1 year: 0.74, 0.68, and 0.80; P = .499).

Secondary endpoint: description of the procedure-related length of stay

Mean LOS at the hospital setting was limited with a mean stay of 5.26 days (± 4.05), and only 10 patients (13.2%) required intensive care, 5 of whom (6.6% of the overall sample) remained at the ICU setting ≤ 1 day, 3 (3.9%) for 2 days, and 2 (2.6%) for 3 days. In figure 3, the ICU and hospital stays are shown based on the patients' baseline risk (very low in all subgroups).

Figure 3. Length of stay (the hospital and the ICU settings) according to STS PROM score; n = 76 for LOS at the hospital setting, and n = 73 for LOS at the ICU setting. ICU, intensive care unit; LOS, length of stay; STS, Society of Thoracic Surgeons. 

Limitations

Inherently to the nature of this observational study, our findings are subject to a few limitations. First, our findings come from the experience of a single center in a tertiary referral hospital very familiar with the procedure so the extrapolation of these results might be affected by the experience of the heart team. Importantly, sample size, especially in the high-risk group, was limited and subject to high dispersion of values and missing data at follow-up. Therefore, further research with a larger number of patients stratified by risk should confirm our findings. Despite this limitation, we should mention that our results observed in the intermediate-risk group are similar to those reported with larger samples allowing for extensive propensity score cohort adjustment.²⁴

Also, in this analysis, we decided to use the STS score as a measure of patient-risk characterization because, even when we acknowledge that this was originally developed for surgical aortic valve replacement, it is a valid prognostic measure of mortality and occurrence of major complications in TAVI making up a comprehensive assessment of the patients' health status.¹³,²⁶ However, as it has been described in recent research,²⁷,²⁸ it would also be helpful to include a measure of frailty to complete thedjusttments. Unfortunately, this information was not routinely collected through a standard form as part of the standard clinical management when this study was conducted. Nevertheless, we should mention that in our center, the heart team involved in the decision-making process regarding the individualized management of the patients' clinical condition, always considers frailty as a key parameter to better adjust the provision of care. Also, precisely due to this preoperative assessment most patients classified as low risk by the STS score (42.5%) were treated with TAVI instead of open surgery. Hence, frailty is a core aspect of this process, always among other important factors like patient preference, history of chest radiation, previous coronary artery bypass graft or porcelain aorta, and others. Although through the comprehensive analysis of clinical and patient-reported outcomes a consistent and positive tendency in outcomes has been shown, we should mention that our results come from a cohort of patients treated in 2018. Consequently, it would be interesting to conduct a new study in multiple centers to obtain data to compare the evolution of the current clinical practice outcomes to those from 2018. Therefore, further research is warranted to continue this monitorization of outcomes in larger samples of patients.