Outcomes of transcatheter aortic valve implantation for intermediate-risk patients in Australia: the SOLACE-AU trial

Abstract

Background: Few studies have examined the safety and efficacy of transcatheter aortic valve implantation (TAVI) for severe aortic stenosis (AS) in Australian patients. SOLACE-AU was a single-arm, open-label clinical trial conducted in Australian hospitals to determine the safety, performance, and cost implications of TAVI in patients with severe, symptomatic AS at intermediate surgical risk.

Methods: This was a prospective, pragmatic, single-arm, multi-center, observational trial of 199 patients with severe, symptomatic AS treated with transfemoral TAVI using the SAPIEN XT transcatheter heart valve (THV) at 11 hospitals in Australia. The primary outcome was Valve Academic Research Consortium-2 (VARC-2) criteria – a composite of seven safety endpoints. Kaplan–Meier (KM) estimates and descriptive analyses were used to evaluate the impact of transfemoral TAVI on safety and valve performance. We also evaluated patient health-related quality of life (QoL) and healthcare resources used throughout the trial.

Results: The valve was successfully implanted in 88% of patients. The VARC-2 outcome at 30 days was 12.1% (95% CI: 8.3–17.5%), and almost 90% of patients had improved heart failure symptoms at 1 year based on New York Heart Association functional class criteria. Patient QoL remained stable over time, with mean EQ-5D-3L scores being 0.71 ± 0.20 at baseline and 0.71 ± 0.21 at 2 years. Duration of post-procedure hospitalization (mean: 6.9 ± 4.7 days) decreased as procedural familiarity increased. The median total cost of TAVI decreased 10.1% at 3 years after introduction of the procedure at the sites.

Conclusions: The SOLACE-AU trial demonstrated favorable safety and performance of the SAPIEN XT valve in patients with AS at intermediate risk of surgical complications.

Keywords:

• Aortic valve replacement
• cardiothoracic surgery
• resource use
• economic evaluation
• TAVI

Introduction

Aortic stenosis (AS) is the progressive narrowing of the aortic valve^1,2. The ensuing hemodynamic changes further contribute to valvular calcification, stenosis and left-ventricular hypertrophy, worsening cardiac output, and ultimately leading to patient death². Patients with AS are often asymptomatic, despite significant stenosis. However, upon disease progression, aortic stenosis is associated with high morbidity and mortality; the expected mortality rate is 50% for untreated, symptomatic patients³. In Australia, the prevalence of AS in the general population is estimated to range between 2.9% and 13.6%⁴. Of this prevalent population, 21.6% were estimated to have severe AS, with the majority (71.1%) of the severe AS population being symptomatic⁴.

Current treatment options for patients with severe, symptomatic AS involve aortic valve replacement, either through surgical or transcatheter methods. For decades, surgical aortic valve replacement (AVR), in which the diseased valve is replaced by opening the sternum to access the heart, excise the diseased valve and suture a new valve in place, was the only treatment option [4,5]. In AVR, the diseased valve may be replaced with biological or mechanical valve prostheses [5,6]. However, advancing age and comorbidities may preclude patients from AVR, with at least one-third of patients being ineligible for surgery⁷. Patients are considered inoperable or at ultra-high risk of surgical complications upon assessment of key patient co-morbidities by a heart team; this assessment is facilitated through the determination of the Society of Thoracic Surgeons (STS) score for predicting patient mortality. Patients with STS scores predicting 30-day mortality >8% or risk of major morbidity >50% may be considered as high risk⁷. Other considerations in the assessment of patient risk include the degree of aortic calcification, patient frailty and cirrhosis of the liver^5,7. Patients determined to be at significant risk of complications from AVR may be eligible for transcatheter aortic valve implantation (TAVI) in Australia⁸. Transcatheter aortic valve implantation is a minimally invasive intervention involving the catheter-based implantation of a prosthetic valve through either the transfemoral or the transapical approach, within the diseased native aortic valve⁸. Transcatheter bioprosthetic valves may be balloon-expandable or self-expanding. Previous trials have shown that TAVI is non-inferior to AVR in patients at high risk for surgical complications, including operative death, and superior to medical management^3,9–11.

The SAPIEN XT transcatheter heart valve (THV; Edwards Lifesciences; Irvine, CA) is a second-generation TAVI system. It was developed to reduce the incidence of complications associated with TAVI, including stroke and vascular access-site bleeding^9,12. Globally, improved transcatheter valve technologies and increasing operator experience have led to the selection of patients of intermediate and low risk for surgical complications for TAVI^7,12. Currently, there is a Class IIa recommendation for TAVI for patients at intermediate surgical risk, based on the American College of Cardiology/American Heart Association (ACC/AHA) guidelines¹³. Furthermore, the European Society of Cardiology/European Association for Cardio-Thoracic Surgery (EACTS) guidelines favor TAVI over AVR for elderly patients with suitable transfemoral access and elevated surgical risk⁵. These recommendations were based on the increasing evidence for the safety and efficacy of TAVI amongst a population at an intermediate level of risk of AVR complications^7,12. Importantly, recent studies comparing TAVI with AVR in a low-risk population setting (STS <4%) have demonstrated that TAVI is at least non-inferior to AVR^14,15. Hence, it is likely that guidelines for the recommendation of TAVI will be extended to include the low-risk AS population.

The Safety, Performance, Quality of Life and Cost Effectiveness Outcomes of the Edwards SAPIEN XT Transcatheter Heart Valve in an Australian Population (SOLACE-AU) study was conducted to determine the safety and performance of the SAPIEN XT THV in intermediate-risk patients with severe, symptomatic AS. This study also explored changes in quality of life (QoL) post-TAVI, and health care resources usage compared with matched AVR patients in an Australian healthcare environment.

Methods

Study design and oversight

SOLACE-AU was a prospective, pragmatic, single-arm, multi-center, observational trial (NCT01675596) of SAPIEN XT delivered via transfemoral artery access. The trial was designed and monitored by the sponsor (Edwards Lifesciences) in conjunction with a steering committee composed of the principal investigator, selected co-investigators, and other clinical trial experts. The steering committee met regularly to monitor patient enrollment and site compliance. An independent clinical events committee (CEC) was also appointed for the adjudication of select outcomes. The CEC was composed of three interventional cardiologists who were not investigators in the trial and had no financial relationship with the sponsor or the clinical sites. A human research ethics committee approved the study protocol before study activities commenced. The manuscript was prepared in accord with the Strengthening of Reporting of Observational Studies in Epidemiology (STROBE) guidelines¹⁶.

Patients

From April 2012 to February 2016, patients with severe, symptomatic AS were enrolled across 11 sites in Australia. Eligible patients had senile, degenerative AS with echocardiographically derived criteria. Symptomatic AS was demonstrated by the New York Heart Association (NYHA) functional class II or greater. Patients with severe, symptomatic AS were assessed by a multidisciplinary heart team and considered eligible for the study if they had an intermediate risk of mortality within 30 days of AVR. An intermediate risk of surgical complications was defined based on patient age ≥70 years and an STS-predicted risk of mortality (PROM) score ≥4. Patients with an STS <4 were eligible for the study if other clinical characteristics, such as advanced liver disease or complications with prior cardiac surgery, indicated an intermediate risk of complications from AVR. If concomitant coronary disease was present, the heart team had to agree on the treatment strategy.

Patients were excluded if they were <70 years of age, or had a native aortic annulus <18 mm or >25 mm. Additional exclusion criteria included the presence of congenital aortic stenosis, unicuspid or bicuspid aortic valves, non-calcific acquired AS, mixed aortic valve disease, and myocardial infarction (MI) within 1 month of treatment, and inability to tolerate anticoagulation therapy. (Full inclusion and exclusion criteria are presented in Supplementary Appendix A).

Patients who received transfemoral SAPIEN XT were followed for a minimum of 2 years and a maximum of 5 years.

Device and procedure

The balloon-expandable SAPIEN XT system and the TAVI procedure have been previously described^12,17. Valve placement was through the transfemoral artery, using a 23, 26, or 29 mm THV. All patients received ≥100 mg aspirin, and low molecular weight or unfractionated heparin, pre-procedure. Post-procedure, patients received ≥100 mg aspirin daily, and clopidogrel, warfarin, or dabigatran for a minimum of 1 month.

Study outcomes

Primary outcome

The primary outcome for SOLACE-AU was a composite of 7 safety variables collected 30 days post-index procedure, based on Valve Academic Research Consortium-2 (VARC-2) criteria¹⁸. The safety variables were all-cause mortality, all stroke, life-threatening bleeding, acute kidney injury (stage 3), coronary artery obstruction requiring intervention, major vascular complications, and valve-related dysfunction requiring a repeat procedure.

Secondary outcomes

Key secondary safety and performance outcomes were evaluated throughout the SOLACE-AU trial at 30 days, 6 months, and annually up to 2 years post-index procedure. The outcomes were the rate of all-cause mortality, cardiovascular mortality, stroke, vascular complications, acute kidney injury, device success, and NYHA functional class changes.

Patient quality of life and resource use

Health-related QoL was assessed using both the short form-36 version 2 (SF-36v2) and EuroQol-5-dimension-3-level (EQ-5D-3L) questionnaires^19,20 at 30 days, and 1 and 2 years post-procedure, at the minimum. The SF-36v2 is a self-reported, 36-item questionnaire that assesses patient QoL across 35 items for eight scales of physical and mental health. The aggregated physical health component score (PCS) and the mental health component score (MCS) may be estimated when scores for all of the eight domains are available.

The EQ-5D-3L is a multi-attribute utility instrument for assessing the overall health state of an individual²⁰. The questionnaire encompasses five dimensions of health (mobility, self-care, usual activities, pain or discomfort, and anxiety or depression). Three levels of responses, ranging from no problems to severe problems, are possible for each dimension in the EQ-5D-3L. The scores were converted into patient utility scores using Australian population-specific weights²⁰. Utility values range in scale from 0–1, with 1 representing full health and 0 representing death²¹. In economic evaluation studies, utility values are used to weight the clinical benefits of interventions (survival) by the QoL associated with post-procedure health²¹.

Additionally, data pertaining to the number of admissions and total days of stay in hospital at 12 months post-index procedure were collected as secondary outcomes, as were data on patient emergency department (ED) attendances and outpatient visits. The rate of permanent pacemaker (PPM) implantation was collected as part of the key safety outcomes. The incidence of new PPM implantation was included in the health care resource use analysis, to determine its impact on health resources use^2,22–25.

All data were site-reported and monitored with independent adjudication of serious adverse events. All adverse events were reclassified according to VARC-2 criteria after these definitions became available. There were no core laboratories used.

Statistical analyses

Data on patient QoL, duration of hospitalization, and resources usage were presented through descriptive and summary statistics unless otherwise specified. For continuous variables, mean (±standard deviation [SD]) or median (interquartile range [IQR]) were presented as descriptive statistics. For binary variables, proportions or rates were presented as descriptive statistics.

Two populations were analyzed. The as-treated population (AT) comprised all enrolled patients who started the procedure, whether it was completed or not. This population was analyzed for the safety of SAPIEN XT. The valve implant (VI) population was a subset of the AT population analyzed for SAPIEN XT performance. Patients in the VI population completed the SAPIEN XT implant procedure.

Patient safety and performance analyses

The primary outcome was analyzed using Kaplan Meier (KM) event rates and the Greenwood standard error; SAS defaults were used for confidence bands and transformations on survival time for the confidence limits. Secondary outcomes, except for NYHA functional class, device success, patient length of stay (LOS) in the hospital, health care resources usage and QoL, were also analyzed using KM estimates.

Patient quality of life and resource use analyses

The impacts of the site and procedural familiarity with TAVI on patient length of stay (LOS) in hospital were also explored. Median, IQR, lower and upper quartiles, and outlier values for patient LOS were estimated over each year from site introduction of TAVI to describe changes in patient LOS across 11 sites over time. Furthermore, patient LOS was stratified based on the number of procedures performed during the trial to explore the impact of operator familiarity with TAVI on LOS and hospital resources use. Sites performing ≤20 procedures were considered to have performed a low volume of TAVI cases, based on Australian guidance for operator and institutional requirements for a TAVI program⁸.

All statistical analyses were performed with the use of STATA software, version 14 (StataCorp, College Station, TX) and SAS version 9.4 (SAS Institute, Cary, NC).

Cost analyses

At the time of this report, there was no Australian Refined Diagnosis Related Group (AR-DRG) code listed for TAVI. As AR-DRGs are used in the estimation of health care resources usage and health care costs, an exploratory analysis was performed using hospital ‘cost-bucket’ data and data from the National Hospital Cost Data Collection (NHCDC) reports. As this represents an exploratory analysis, the methodology used in the estimation of the cost impacts of TAVI for patients with severe, symptomatic AS at intermediate surgical risk is presented in Supplementary Appendix C.

Results

Patient characteristics and procedural outcomes

A total of 199 patients were enrolled in the SOLACE-AU trial. Of these patients, 15 withdrew consent, one patient was lost-to-follow-up and eight patients exited the study for other reasons. A total of 88.7% of patients had completed the 2-year follow-up visit, and the mean follow-up time was 2.7 ± 1.34 years. The AT population comprised of 199 patients, and 198 patients completed the implantation procedure (VI population). Table 1 summarises the baseline patient characteristics and procedural characteristics for the AT population.

Table 1. Patient characteristics at baseline (all treated population).

Characteristic	TAVI (n = 199)
Age, years, mean ± SD	85.5 ± 4.51
Gender, n (%)
Male	91 (45.7)
Female	108 (54.3)
Height, cm, mean ± SD	163.4 ± 9.31
Weight, kg, mean ± SD	72.7 ± 15.63
Body mass index, kg/m^2, mean ± SD	27.2 ± 5.10
Body surface area, m^2, mean ± SD	1.8 ± 0.21
Risk scores, mean ± SD
STS score	5.9 ± 2.59
Logistic EuroSCORE	18.5 ± 11.19
EuroSCORE	6.1 ± 6.06
NYHA functional class, n (%)
Class I	0
Class II	43 (21.6)
Class III	143 (71.9)
Class IV	13 (6.5)
Grouped NYHA functional class, n (%)
Class I/II	43 (21.6)
Class III/IV	156 (78.4)
Co-morbidities, n (%)
Congestive heart failure	93 (46.7)
Coronary artery disease	128 (64.3)
Hypertension	153 (76.9)
Carotid disease	29 (14.6)
Peripheral vascular disease	27 (13.6)
Myocardial infarction	36 (18.1)
Cerebrovascular accident	21 (10.6)
Transient ischemic attack	13 (6.5)
Percutaneous coronary intervention	46 (23.1)
Coronary artery bypass graft	44 (22.1)
Atrial fibrillation	72 (36.2)
Pacemaker	22 (11.1)
COPD	40 (20.1)
Oxygen-dependent COPD	1 (2.5)
Renal insufficiency (Creatinin ≥2 mg/dL)	10 (5.0)
Diabetes	51 (25.6)
Procedural outcomes
Total cath lab or operating room time, min, mean ± SD	164.3 ± 47.39
Total procedure time, min, mean ± SD	75.8 ± 34.31
Anesthesia, n (%)
Conscious sedation	24 (12.1)
General	175 (87.9)
Fluoroscopy time (min, mean ± SD)	19.1 ± 8.00
Conversion to AVR, n	0
Device malfunction, n	0

Abbreviations. COPD, chronic obstructive pulmonary disease; min, minutes; NYHA, New York Heart Association; AVR, aortic valve replacement; SD, standard deviation; STS, Society of Thoracic Surgeons; TAVI, transcatheter aortic valve implantation.

The mean patient age was 86 years, and there were slightly more females (54%) in the cohort. Common patient comorbidities included hypertension (76.9%), coronary artery disease (64.3%), congestive heart failure (46.7%), atrial fibrillation (36.2%), previous coronary artery bypass grafting (22.1%) and prior percutaneous coronary intervention (20.5%). The mean procedure time was 76 min, and there were no device malfunctions. No patients required conversion from TAVI to surgery.

Patient safety outcomes

The primary and key secondary safety outcomes of the SOLACE-AU trial are summarised below (see Figure 1 and Table 2). Based on the KM analyses of the VARC-2 composite outcome, the primary outcome safety variables (all-cause mortality, stroke and other key safety criteria) occurred among 12.1% (95% CI: 8.3–17.5%) of patients in the AT population within 30 days (see Figure 1). Of the VARC-2 outcomes, patients had experienced major vascular complications (6.0%), bleeding (3.5%), and stroke (3.5%) within the first 30 days (see Supplementary Appendix B). The occurrence of key secondary safety events throughout the study was low, with the most commonly observed safety outcome at 30 days being vascular complications (KM estimate: 13.1%; 95% CI: 9.1–18.6%) (see Table 2).

Figure 1. Primary endpoint through 30 days (AT population).

Table 2. Key primary and secondary outcomes (AT population).

Outcome, KM estimate, [95% CI]	TAVI (n = 199)
	30 days	6 months^a	1 year^a
Primary outcome^b	12.1 [8.3, 17.5]	NE	NE
Secondary safety outcomes
All-cause mortality	2.5 [1.1, 5.9]	6.1 [3.5, 10.5]	8.7 [5.5, 13.6]
Cardiovascular mortality	2.5 [1.1, 5.9]	4.1 [2.0, 7.9]	4.6 [2.4, 8.6]
Stroke	3.5 [1.7, 7.3]	3.1 [1.4, 6.7]	4.7 [2.5, 8.9]
Vascular complications	13.1 [9.1, 18.6]	NE	NE
Acute kidney injury^c	3.0 [1.4, 6.6]	NE	NE

Abbreviations. CI, confidence interval; KM, Kaplan-Meier; NE, not evaluated; TAVI, transcatheter aortic valve implantation. ^aEndpoints at 6-months and 1 year are site-reported and have not been adjudicated by the CEC. ^bThe primary outcome was a composite of: all-cause mortality, all-stroke, life-threatening bleeding, acute kidney injury, coronary artery obstruction requiring intervention, major vascular complication and valve-related dysfunction requiring repeat procedures. ^cThe secondary outcome of acute kidney injury was assessed at 7 days.

The KM curve for all-cause mortality through 2 years for the AT population is presented in Figure 2. Over a period of 2 years, 32 deaths (16.8%) occurred. Of these deaths, 16 patients (8.8%) died due to cardiac causes, 14 (7.8%) due to non-cardiac causes, and two (1.1%) died due to unknown causes.

Figure 2. Site-reported all-cause mortality through 2 years (AT population).

Patient performance outcomes

Table 3 summarises the trends in patient NYHA functional class and valve hemodynamics over time observed for the VI population. A significant proportion of patients experienced improvement in NYHA functional class at 30 days, which was maintained to 1 year post-implant. Furthermore, the proportion of individuals with NYHA Class III declined markedly from 71.7% at baseline, to 6.7% at 1 year post-implantation. Only a small proportion (6.3%) of patients had moderate-to-severe paravalvular leakage (PVL) at baseline, and the majority of patients had mild or less PVL throughout the study. Declines in mean (mean decline: 39.5 mmHg) and peak (mean decline: 61.7 mmHg) aortic valve gradients, along with increases to valve effective orifice area (EOA) (mean increase: 1.0 cm², SD: 0.43 cm²) were observed between baseline and 30 days, and were maintained over 2 years of follow-up (Supplementary Appendix B).

Table 3. NYHA functional class and valve haemodynamics over time (VI population).

Outcome	Visit
NYHA functional class (%n)	Baseline (n = 198)	30 Days (n = 186)	6 Months (n = 168)	1 Year (n = 149)
Class I	0	97 (52.2)	97 (57.7)	83 (55.7)
Class II	43 (21.7)	72 (38.7)	62 (36.9)	56 (37.6)
Class III	142 (71.7)	16 (8.6)	7 (4.2)	10 (6.7)
Class IV	13 (6.6)	1 (0.5)	2 (1.2)	0 (0.0)
PVL	Post-procedure^a(n = 176)	30 days (n = 178)	1 year (n = 134)	2 years (n = 106)
Grade
None	41 (23.3)	32 (18.0)	37 (27.6)	30 (28.3)
Trace	43 (24.4)	49 (27.5)	32 (23.9)	19 (17.9)
Mild	81 (46.0)	77 (43.3)	51 (38.1)	47 (44.3)
Moderate	10 (5.7)	20 (11.2)	14 (10.4)	10 (9.4)
Severe	1 (0.6)	0	0	0

Abbreviations. NYHA, New York Heart Association; PVL, paravalvular leak. ^aParavalvular leak was assessed post-procedure.

Quality of life

Table 4 summarises patient QoL over time. Standardized population norms for both the mental health score and the physical component score for the SF-36v2 are 50³. For the VI population, the mean SF-36v2 physical (PCS) and mental component scores (MCS) were 36.1 ± 8.5 and 47.9 ± 10.9, respectively, at baseline. These scores improved slightly at 30 days post-procedure, but the differences were not clinically significant. Additionally, these scores did not vary significantly at subsequent follow-ups.

Table 4. Patient health-related QoL over time (VI population).

QoL measure	Visit
	Baseline	30 days	6 months	1 year	2 years
SF-36v2 Component score, mean (SD)
PCS	36.1 (8.5)	39.2 (8.3)	39.7 (9.3)	39.5 (8.4)	38.1 (9.6)
MCS	47.9 (10.9)	50.8 (11.1)	50.6 (10.3)	51.9 (9.6)	51.7 (10.3)
EQ-5D-3L Domains (m, %n)	n = 194	n = 189	n = 169	n = 147	n = 119
Mobility
1 = no problem	69 (35.6)	85 (45.0)	70 (41.4)	49 (33.3)	39 (32.8)
2 = some problem	125 (64.4)	104 (55.0)	98 (58.0)	97 (66.0)	79 (66.4)
3 = confined to bed	0	0	1 (0.6)	1 (0.7)	1 (0.8)
Self-care, n (%)
1 = no problem	154 (79.4)	156 (83.0)	140 (82.8)	114 (77.6)	92 (77.3)
2 = some problem	37 (19.1)	30 (16.0)	27 (16.0)	33 (22.4)	25 (21.0)
3 = unable	3 (1.5)	2 (1.1)	2 (1.2)	0	2 (1.7)
Usual activity, n (%)
1 = no problem	74 (38.1)	84 (44.7)	81 (48.2)	59 (40.1)	47 (39.5)
2 = some problem	106 (54.6)	91 (48.4)	79 (47.0)	82 (55.8)	64 (53.8)
3 = unable	14 (7.2)	13 (6.9)	8 (4.8)	6 (4.1)	8 (6.7)
Pain/discomfort, n (%)
1 = no	83 (42.8)	99 (52.7)	83 (49.4)	57 (38.8)	47 (39.5)
2 = moderate	101 (52.1)	82 (43.6)	80 (47.6)	83 (56.5)	65 (54.6)
3 = extreme	10 (5.2)	7 (3.7)	5 (3.0)	7 (4.8)	7 (5.9)
Anxiety/depression, n (%)
1 = no	124 (63.9)	129 (68.6)	112 (66.7)	108 (73.5)	89 (74.8)
2 = moderate	66 (34.0)	59 (31.4)	55 (32.7)	38 (25.9)	28 (23.5)
3 = extreme	4 (2.1)	0	1 (0.6)	1 (0.7)	2 (1.7)
EQ-5D-3L Utility Score, mean (SD)	0.71 (0.20)	0.75 (0.19)	0.75 (0.19)	0.72 (0.19)	0.71 (0.21)

Abbreviations. EQ-5D-3L, EuroQol-5-dimension-3-level; MCS, mental component score; PCS, physical component score; QoL, quality-of-life; SD, standard deviation component score; SF-36v2, short-form 36 version 2.

Similarly, patient health-related QoL, as measured using the EQ-5D-3L, remained relatively stable over time. The mean patient utility score was 0.71 at baseline, improving slightly to 0.75 at the 30-day, post-procedure follow-up, and decreasing slightly to 0.71 at 2 years (see Table 4)²⁰.

Length of stay and resource usage analyses

Figure 3 presents the hospital LOS by the number of years of experience in performing TAVI and the number of TAVI procedures performed. In the first year post TAVI, the average LOS was 6.7 ± 2.9 days. In the second year, the average LOS increased to 7.4 ± 6.2 days. In the third year, the average LOS decreased to 6.6 ± 3.6 days, and then to 5.4 ± 3.0 in year 4. Of the 11 sites that performed TAVI, eight were considered to perform a low volume of procedures (≤20 TAVI); the average LOS was 8.4 ± 6.4 days for these sites. For sites performing fewer than 10 procedures, the average hospital LOS was 9.0 ± 7.7 days compared with an average of 6.0 ± 3.2 days for sites performing more than 30 TAVI procedures.

Figure 3. Length-of-stay over time, and procedural experience (VI population).

The results of the exploratory costing analyses are presented in Supplementary Appendix C. Briefly, based on micro-cost data obtained from the major study sites, the median total cost of TAVI in year 1 was $61,642 (IQR: $50,014–$70,077), decreasing to $55,400 (IQR: $49,647–$60,162) in year 4 for the Fiona Stanley Hospital (FSH) and the Royal Perth Hospital (RPH). TAVI procedural costs remained relatively stable over time. The average cost of a prosthesis was $29,574 ± $5,514 and varied across the major sites and over the four years of the trial depending on price-volume-service agreements with Edwards Lifesciences.

The number of ED attendances and outpatient visits up to 1 year post-procedure are presented in Table 5. The cumulative number of ED attendances were heavily skewed for both non-admitted (median attendances: 2; IQR: 1–2; range: 1–5) and admitted patients (median: 3; IQR: 2–4; range: 1–9) at the 1-year, follow-up period. The cumulative number of outpatient visits per patient were also heavily skewed at the 1-year follow-up period (median: 3; IQR: 3–5; range: 1–10). Costs attributed to ED attendances and outpatient visits were correspondingly skewed; at the 1-year, follow-up period, the median cumulative cost of ED attendances for non-admittted patients was $540 (IQR: $483, $1,029), and $2,074 (IQR: $1,041–$3,072) for admitted ED patients (see Supplementary Appendix C). The median cumulative cost of outpatient visits at the 1-year, follow-up period, after allowing for estimates in temporal trends, varied from $859 (IQR: $859–$1,432) to $1,163 (IQR: $1,163–$1,938).

Table 5. ED attendances and outpatient visits over time (VI population).

Variable	30-day visit	6-month visit	1-year visit
ED attendances
Cumulative ED attendances^a per patient
Non-admitted patients^b
Mean (SD) [range]	1.2 (0.4) [1, 2]	1.6 (1.0) [1, 5]	2.0 (1.2) [1, 5]
Median [IQR]	1.0 [1.0, 1.0]	1.0 [1.0, 2.0]	2.0 [1.0, 2.0]
Admitted patients^c
Mean (SD) [range]	1.4 (0.8) [1, 3]	2.2 (1.6) [1, 7]	3.3 (1.9) [1, 9]
Median [IQR]	1.0 [1.0, 1.0]	2 [1.0, 2.0]	3.0 [2.0, 4.0]
Outpatient visits
Cumulative visits per patient^a
Mean (SD) [range]	1.2 (0.45) [1, 3]	2.8 (1.3) [1, 9]	3.7 (1.7) [1, 10]
Median [IQR]	1.0 [1.0, 1.0]	2.0 [2.0, 3.0]	3.0 [3.0, 5.0]

Abbreviations. ED, emergency department; IQR, interquartile range. ^aCumulative mean visits per patient or outpatient at 0–30 days, 0 days to 6 months and 0 days to 1 year. ^bNon-admitted patients include patients discharged to home, nursing home or died. ^cPatients discharged to medical ward, ICU, CCU, surgical ward or operation theatre.

Discussion

Valve safety and performance

Following the success of the PARTNER 2 A trial that demonstrated the safety and efficacy of TAVI using balloon-expandable THV in patients with severe AS at intermediate operative risk¹², the SOLACE-AU trial determined the initial experience of TAVI using the second-generation SAPIEN XT THV for similar patients in Australia. It should be noted that the third-generation Sapien 3 valve is now approved for use in Australia⁷. Based on the primary safety outcome, the majority of patients experienced no significant safety event at 30 days post-procedure. The safety of TAVI using the SAPIEN XT for this intermediate-risk patient population was further supported by KM analyses of key secondary safety outcomes.

Data pertaining to patient NYHA class and valve hemodynamic characteristics were considered in the evaluation of TAVI efficacy. Overall, the rate of successful device implantation was high, and the majority of patients saw improvement in NYHA functional class at 30 days post-procedure. Improvements to patient NYHA functional class were maintained through 1 year following implantation; importantly, the proportion of patients with NYHA functional Class III heart failure symptoms in the cohort declined drastically at 30 days post-procedure. The frequency and rate of PVL were broadly comparable with data from the PARTNER 2A trial, with the majority of patients having only mild PVL over time¹².

It should be noted that recent trials have demonstrated that TAVI was at least non-inferior to AVR for patients at low risk of surgical complications (STS-PROM <4)^14,15. In the recent PARTNER 3 trial comparing TAVI using a balloon-expanded prosthesis with AVR, the rate of occurrence of the composite outcome of all-cause death, stroke, or rehospitalization was considerably lower for patients treated with TAVI relative to patients treated using AVR at 1 year (hazard ratio (HR): 0.54; 95% CI: 0.37–0.79) (p = .001)¹⁴. A separate trial comparing TAVI using a self-expanding prosthesis with AVR, concluded that TAVI was non-inferior based on the reduction in composite outcome of all-cause death or disabling stroke at 24 months¹⁵. Such promising results highlight the potential utility in expanding guidelines for TAVI to include the AS population at low surgical risk of complications in Australia.

Patient quality of life and length of stay

Slight variations around patient QoL over time were observed. Based on the EQ-5D-3L, there was a numerical improvement to patient QoL at 30 days post-procedure compared with baseline estimates, and a gradual return to baseline levels over 2 years. It is expected that elderly patients (mean age of 86 years) would have a background decrease in quality of life. The results from the EQ-5D-3L provide support that TAVI improves QoL as patients had a similar EQ-5D-3L score at 2 years follow-up as they had a baseline (i.e. an initial gain that declined over time). Without TAVI these patients likely would have died or had substantially lower QoL. However, these variations were not clinically significant, and overall, patients QoL remained stable over time. Similarly, PCS and MCS scores measured using the SF-36v2 questionnaire varied slightly through time, and overall, reflect the stable patient QoL measured using the EQ-5D-3L. It should be noted that utility values estimated using the SF-36v2, through conversion to the short form, 6-dimension questionnaire, may not be appropriate for the severe, symptomatic AS population. This is based on a prior study that found that the SF-6D was less discriminative in the setting of severe cardiovascular disease conditions²¹.

Concerning patient index hospitalizations, the average hospital LOS for patients in the SOLACE-AU cohort was favorable in comparison with the initial reports from the PARTNER 2 trials, where the average LOS was 8 days^3,10. Furthermore, the results of the SOLACE-AU study demonstrated that as operators become more experienced (i.e. more procedures performed and more years of experience), the average LOS markedly decreased (6.0 days from 30+ procedures), whereas those with low throughput (<10 procedures) had an average LOS of 9.0 ± 7.7 days. Operators with low throughput were the new adopters of TAVI and becoming familiar with the procedure. This familiarity is not just a training matter for the interventionalist, but requires a major shift in management throughout the hospital care pathway for TAVI patients. For example, the nursing staff learn how to manage TAVI patients; likewise coronary care staff may be extra cautious in managing TAVI patients and be more vigilant with these patients, detaining them longer. Reduced LOS increases bed availability and enables greater throughput of patients in a hospital and a subsequent reduction in hospital resources utilization. Importantly, based on data from the PARTNER 3 trial, the median length of index hospitalization was 3.0 (IQR: 2.0–3.0). Hence, further reductions in resource use may be possible upon the introduction of the latest generation of balloon-expandable devices, and the inclusion of low-risk AS patients for TAVI in Australian guidelines¹⁴.

Healthcare resources usage

In the exploratory costing analyses, it was noted that the key driver of index hospitalization costs was the prosthesis itself. While minor variations to prosthesis costs were noted over time and across hospitals, prosthesis costs drove 51.0% of the costs of index hospitalisation (Supplementary Appendix C). Costs associated with the index TAVI procedure decreased between the first and fourth year of introducing the TAVI procedure to TPCH, the FSH and the RPH; this was in line with decreasing LOS at these sites. The mean cumulative cost of outpatient visits per patient and the cumulative cost of ED attendances at the 1 year follow-up were heavily skewed. This was attributed to the large degree of variance around the number of ED attendances per patient for a given period. Furthermore, healthcare costs are typically positively skewed, with the majority of costs being attributed to a minority of individuals²⁴.

As part of this analysis on health resources usage, the rate of PPM implantation was explored (see Supplementary Appendix D). The incidence of PPM implantation in the SOLACE-AU trial was consistent with other trials evaluating TAVI using balloon-expanding prostheses^9,11,12. Further, in comparison with TAVI studies using self-expanding prostheses, patients in the SOLACE-AU trial had markedly lower rates of PPM implantation^22,25. This was in line with literature comparing self-expanding and balloon-expandable prostheses, and highlights the potential cost-savings associated with the use of balloon-expandable prostheses^23,25.

Limitations

This study had some limitations. Firstly, this was an open-label, non-randomized study without a comparator arm, with patients selected for a procedure following an assessment of risk for AVR. Hence, it was not possible to assess the performance and safety of transfemoral TAVI compared with AVR or transapical TAVI. It should be noted that the comparisons between SOLACE-AU and current literature on AVR were indirect in nature, and should be interpreted in this light. Furthermore, the number of patients recruited across hospitals over time was small with half of the patient population being recruited by three hospitals in Australia (TPCH/FSH/RPH). Patient recruitment across other sites contributed to 2.0% (Royal North Shore) to 24% (Alfred Hospital) of the total patient cohort.

The estimates pertaining to health care resource use were exploratory. Furthermore, the accuracy of costs estimated for ED attendances and outpatient visits in the analyses are limited by the application of a general, nationwide cost per admission from NHCDC data. A key strength of this study is the bottom-up, micro-costing analyses using hospital cost data from TPCH and FSH/RPH is that the direct costs attributed to a procedure may be estimated²⁶. Importantly, micro-cost data from TPCH and FSH/RPH comprised 99 patients (as-treated population) and represents approximately 50% of the SOLACE-AU as-treated cohort. Further, follow-up data was available up to 2 years from procedure, allowing for the assessment of long-term safety and performance of TAVI.

Conclusions

The SOLACE-AU study demonstrated that the TAVI procedure for intermediate-risk patients was associated with a favorable safety and device performance profile. Additionally, patient QoL remained stable over time. It is likely that TAVI will lead to a reduction in health care resources usage and costs. However, the cost savings estimated in this study are exploratory and may not be applicable outside of the local setting. Ultimately, a randomized, controlled trial of this procedure compared with AVR within the local clinical context is essential to define the true value of this novel technology in Australia.

Transparency

Declaration of funding

The SOLACE-AU trial was funded by Edwards Lifesciences.

Declaration of financial/other relationships

PL and PS were funded by Edwards Lifesciences to provide this report. JME peer reviewers on this manuscript have received an honorarium from JME for their review work, but have no other relevant financial relationships to disclose. All authors met the ICMJE criteria for authorship.

Acknowledgements

Tracey Fine, MS, ELS of Edwards Lifesciences provided medical writing services.