UK Transcatheter Aortic Valve Implantation trial — time to differentiate “non-inferiority” from “equivalence”?

Abstract

The UK Transcatheter Aortic Valve Implantation (UK-TAVI) trial has compared transcatheter aortic valve implantation (TAVI) and surgical aortic valve replacement (SAVR) in patients above the age of 70 years with moderate operative risk. The primary outcome was 1-year all-cause mortality. This was a non-inferiority trial that concluded that TAVI in this cohort was non-inferior to SAVR. This appraisal reports the main findings of the study and discusses the strengths and weaknesses of the trial.

Introduction

The comparison between transcatheter aortic valve implantation (TAVI) and surgical aortic valve replacement (SAVR) continues with the latest trial coming out of the United Kingdom (UK). The UK Transcatheter Aortic Valve Implantation (UK-TAVI) trial was conducted to compare TAVI with SAVR in patients with severe, symptomatic aortic stenosis [1].

Trial design

The UK-TAVI study was not funded by the industry and was in fact an investigator-initiated multi-centric, randomized control trial that was funded by the National Institute for Health Research. All the 34 centres carrying out TAVI in the UK were included in the trial. The participants included those above the age of 70 years who were deemed to have moderate operative risk. The risk assessment was done by a multidisciplinary team and while the European System for Cardiac Operative Risk Evaluation II (EuroSCORE II) and the Society of Thoracic Surgeons–predicted mortality (STS-PROM) scores were calculated, they did not form the basis for patient selection. Patients above the age of 80 years were deemed eligible irrespective of their comorbid status. This was a non-inferiority trial. The initial sample size was 808 with an inferiority margin of 7.5% based on expected mortality of 12.5% after surgery. However, after an interim analysis of outcomes, which was planned at the start of the study, the sample size was increased to 890 and the inferiority margin reduced to 5%. Intervention included TAVI and SAVR. Participants undergoing TAVI were allowed to have any valve with a Conformité Européenne (CE, meaning European conformity) mark; access route was left to the discretion of the interventionist. Similarly, the choice of valve was left to the surgeon’s discretion except for sutureless valves which were not allowed in the trial. The comparison was carried out by randomizing participants into SAVR and TAVI in 1:1 ratio. Randomization used minimization, which is a technique whereby imbalances in covariates can be minimized. In relation to the trial, stratification was done with respect to age, site, and presence of coronary artery disease. The patients and the site personnel were unblinded. Outcome (primary) was all-cause mortality at 1 year. Secondary outcomes included cardiovascular death, stroke, reintervention, and multiple composites of these variables as well as vascular complications, major bleeding, infective endocarditis, and myocardial infarction. Need for permanent pacing or renal replacement therapy was also included. Follow-up was planned for a minimum of 5 years.

Trial findings

The median age of the participants in the trial was 81 years and more than two-thirds of the participants in each arm were older than 80 years. The representation of female participants was high with 211 (46.1%) in the TAVI arm and 213 (46.8%) in the SAVR arm being females. Coronary artery disease was present in roughly one-third of the patients in both the arms. The overall median EuroSCORE II was 2, which was lower in the age group 70–79 years (TAVI 1.7, SAVR 1.8). Similarly, in those 80 years or above, the median EuroSCORE II was low for both TAVI (2.2) and SAVR (2.2). The STS-PROM score was also low both for 70–79 years age group (TAVI 2.1, SAVR − 2.0) and those ≥ 80 years (TAVI 2.9, SAVR − 2.9). The crossover rate was low with 5 patients crossing over from TAVI to SAVR and 17 patients from SAVR to TAVI.

Primary outcome, which was all-cause mortality at one year, was 21 (4.6%) among the TAVI group and 30 (6.6%) among the SAVR patients. The absolute risk difference in death from any cause at 1 year was 2.0% with the upper limit of the confidence interval (CI) being 1.2 which was lower than the inferiority margin set at 5. As the CI did not cross the inferiority threshold of 5.0, TAVI was proven to be non-inferior to SAVR in the study.

Median length of hospital stay was significantly shorter with TAVI (3 days vs 8 days) and TAVI was superior to SAVR in terms of major bleeding events, but was significantly worse in terms of vascular complications and need for permanent pacing. The 1-year incidence of major bleeding was 7.2% in TAVI vs 20.2% in SAVR, p < 0.001, and the incidence of vascular complications with TAVI was significantly higher at 10.3% vs 2.4% in SAVR, p < 0.001. Need for permanent pacemaker was also significantly higher in the TAVI arm with an incidence of 14.2% as opposed to 7.3% in the SAVR arm, p < 0.001. No differences were observed in the stroke rate either at 30 days or at 1 year. The incidence of mild and moderate aortic regurgitation (AR) was significantly more common with TAVI both at 30 days and at 1 year. At 1 year, mild AR was present in 38.3% of TAVI patients compared to 11.7% in the SAVR group, while moderate AR was seen in 2.3% and 0.6% cases, respectively. At 1 year, there was no difference between TAVI and SAVR with respect to Six-Minute Walk Test, New York Heart Association class, or Minnesota Living with Heart Failure Questionnaire scores. No significant difference was observed between the two interventions with regard to independence in activities of daily living at 1 year.

Strengths of the UK-TAVI trial

The UK-TAVI trial, rather than focussing on only high-volume centres of excellence or high-volume operators with significant experience, included all centres and interventionists where TAVI was carried out in the UK. Age was the prime criteria for inclusion and risk assessment was carried out by a multidisciplinary team which allowed consideration for other important features like frailty [2]. Every centre, where TAVI was carried out in the UK, was part of the study and even included suitable patients with bicuspid valves. Thus, the study was close to being representative of real-life practices. Since the study was not sponsored by industry, the study was not limited by the valve type or access route and potential biases with trial design were avoided.

Weakness of the UK-TAVI trial

The study intended to include patients with moderate or high risk [3]; however, both the STS-PROM and the EuroSCORE II suggest that the patients recruited were in fact low-risk patients. There is also no information on patients screened, and hence it is difficult to truly assess the generalisability or the process of patient selection in the trial. Besides, this was another “non-inferiority” trial and questions remain on the width of the inferiority margin chosen.

Comment

The UK-TAVI is a well-conducted, pragmatic study, free of perceived biases in industry-sponsored trials; however, several issues remain. Is mortality at 1 year a reliable outcome measure for patients undergoing valve replacements? Five-year follow-up data from the PARTNER 2 trial shows a significantly higher incidence of paravalvular AR, repeat hospitalizations, and aortic valve reinterventions in patients undergoing TAVI [4] suggesting that 1-year mortality rates may not provide an adequate benchmark for comparison. The second issue is the inherent weakness of the non-inferiority design and the fact that “non-inferiority” is not the same as “equivalence” [5]. While the authors have stayed true to the term “non-inferior”, the interpretation of these studies, especially in the lay media, is inevitably of “equivalence” which is scientifically inaccurate, irresponsible, and misleading [6] and is perhaps one of the reasons why “non-inferiority” trials are considered to be unethical [7]. Another major misinterpretation of UK-TAVI and other “low-risk” trials [8, 9] comparing TAVI and SAVR, is equating “low-risk” with “younger” patients. It has to be remembered that the minimum age in the UK-TAVI trial was 70 years with the median age being 80 years and the findings of this study should not be extrapolated to “younger” patients with longer life expectancy.

The authors of the UK-TAVI study have to be congratulated for their work. However, the 5-year follow-up results from the trial need to be assessed to confirm “non-inferiority” in the longer term. Moreover, significant amount of caution is required in interpretation and extrapolation of the findings which cannot be used to justify TAVI in patients younger than that included in the trial. Finally, for outcomes of TAVI to be “equivalent” to SAVR, future study design should be of “equivalence” and not of “non-inferiority”.

Funding

None.

Declarations

Ethical approval

Not applicable being a review of published data.

Informed consent

Not applicable as this is a review article.

Conflict of interest

None.