DISCUSSION
The real-world SAVI registries with 500 enrolled patients show very good safety and performance outcomes in patients treated with the ACURATE TATM system. Outcomes across the SAVI-1 and SAVI-2 registries and the ‘TA90’ First-in-man and pilot studies are consistent, e.g. moderate or severe PVL <3.5%, ∼10% permanent pacemaker implantation, and ∼20% 1-year mortality [3, 5]. This is remarkable as the ‘TA 90’ cohort were first-time users, as well as approximately 70% of centres in SAVI-1 [6], and speaks for the ease of use of the device. There was a difference in the rate of conversion to open heart between SAVI-1 and -2, though. This could have been a coincidental finding, but could also reflect the fact that, in general, complication management had improved. Similarly, during the same time period, the GARY registry observed an improvement in technical procedural complications [8]. Notably, the number of valve-in-valve procedures was higher in SAVI-2 (2.4% vs 0.8%)—eventually, patients who would have been converted in the early experience had received a valve-in-valve procedure in SAVI-2. Furthermore, the difference between SAVI-1 and -2 observed for rapid pacing probably reflects the increasing confidence in this technique.
Reasons for low complication rates and ease of use of the device have previously been reported: (i) re-sheathability, (ii) top-down implantation technique compressing the native leaflets and capturing them in the waist of the device, hence reducing the risk of paravalvular leakage and potentially also reducing the risk of coronary occlusion, (iii) stabilization arches allowing a coaxial self-alignment of the valve and stent commissure alignment during fluoroscopy, hence avoiding a stent-post in front of the coronary ostia and potentially easing later interventional access to the coronaries, (iv) the upper crown allows supra-annular anchoring, tactile feedback and stable positioning, (v) the waist conforms to the native annulus, (vi) the lower crown allows for minimal stent protrusion into the left ventricle and hence (in combination with the low radial force needed due to the anatomic shape of the valve) reduces the risk of conduction system interference with subsequent pacemaker implantation, and (vii) the PET skirt acts as a seal to prevent PVL [4–6, 9].
Regarding valve design, there could be a concern that the two-step deployment and the higher need for post-dilatation compared to balloon-expanding valves [10] might lead to a higher stroke rate. With the limitation of site-reported data and the lack of differentiation between minor and major stroke, we did not observe a high stroke rate in the SAVI-registries. This observation is confirmed by a transcranial doppler ultrasound study, comparing the frequency and pattern of high-intensity transient signals (HITs) in 22 patients receiving either the transapical ACURATE TATM or the balloon-expandable SAPIEN XT valve (Edwards Lifesciences LLC, Irvine, CA, USA), which found similar outcomes for both devices [11].
Kempfert et al. [6] compared early SAVI-1 results with those of the FRANCE-2 and SENTINEL registries using the CoreValve (Medtronic Inc., Minneapolis, MN, USA) and SAPIEN/SAPIEN XT prosthesis. Even though they included 75% transfemoral cases, early mortality was comparable, the leakage rate extremely promising, and the pacemaker rate within the range of the SAPIEN system and superior to the CoreValve system. Furthermore, SAVI-1 and -2 outcomes compare well to those of a recent review of transapical second-generation transcatheter heart valves [12], a recently published report of the JUPITER registry using the JenaValve (JenaValve Technology GmbH, Munich, Germany) [13], and the transapical cohort of the GARY registry [14]. Thirty-day mortality in our series was 6.8% compared to 8.9%, 11.1% and 7.7% (in-hospital mortality), and pacemaker were implanted in 10.2% compared to 12.1%, 14.4% and 11.3%, respectively. Bailout situation (valve-in-valve implantations and conversion to open heart) occurred in 2.2% of our series, compared to 5.0% in the JUPITER study and 2.0% (valve-in-valve procedures only) in the GARY registry.
Moderate to severe PVL at 30 days occurred in only 1.9% (2.6% at 1 year) of our cases. Accordingly, in a single-centre series, the ACURATE TATM valve had the lowest PVL-rate of next-generation transapical valves [15], and significantly less PVL than the transfemoral CoreValve prosthesis [16]. Overall, echocardiographic parameters (effective orifice area and mean gradient) and NYHA class remained stable between 30 days and 1 year.
Notably, even though the registries enrolled between 2011 and 2014, and many centres were first-time users, the pacemaker, stroke and PVL rates comply to the ‘future targets for optimal quality centres’ as published in a recent state-of-the-art review which postulated the following limits: new pacemaker <10%, major stroke <2% and moderate to severe PVL <5% [17].
These results confirm that the ACURATE TATM valve adds another valuable option to the TAVI armamentarium. Currently, it is the only commercially available self-expandable valve for transapical access. Due to its unique features, this valve might be particularly useful in (i) patients with low coronary ostia as the tissue will be pulled away from the ostia during implantation [6], (ii) patients with massive calcification, as well as an absence of calcification [5], or even in pure aortic regurgitation [18], (iii) due to its low pacemaker and PVL rates and implantability without rapid pacing—in patients with poor left ventricular ejection fraction, (iv) centres with little experience as the valve is easy to implant and resheathing and repositioning are possible until final release [4–6], and (v) low-volume centres as they ideally restrict themselves to two valve types to allow adequate device experience and therefore should select a valve type that can serve the transfemoral and transapical approach.
Undoubtedly, with the adaption of smaller introducer systems, transfemoral nowadays is the most common access route. Exemplary in Germany, the numbers of transapical cases have remained constant over the years while there has been a massive increase in transfemoral ones [2]. This has probably led to an even more severely diseased patient population for transapical cases, especially related to peripheral artery disease. Correspondingly, SAVI-2 patients had numerically higher risk scores than SAVI-1 patients. Nevertheless, 30-day mortality is nearly identical between both registries and there is no statistically significant difference in 1-year mortality, even though the Kaplan–Meier curves start to diverge after 3 months, which might be indicative of a more severely diseased patient population.
Though the majority of cases are transfemoral nowadays [2], the transapical approach is still relevant. When transfemoral access is not possible, it is the access of choice, as reported in a recently conducted survey [19]. It has not only the advantage of a short distance to the annulus, resulting in precise control of the device, but also avoids crossing of the aortic arch, which is especially helpful in severely calcified anatomy, and can ultimately be applied in almost all patients [10]. Furthermore, there is a development towards devices with smaller transapical delivery systems. Recently, the combination of ACURATE neoTM with a 22F outer diameter transapical delivery system has been successfully tested [4] and a CE approval trial is currently enrolling.
Limitations of the SAVI registries are the ones inherent to registries, such as self-reporting of events. However, risk of under-reporting was minimized due to the application of yes/no questions for safety endpoints. Furthermore, data were not monitored, nor adjudicated, no core laboratory was used, and SAVI-1 did not include VARC recommendations [7]. As—due to the nature of a registry—echocardiographic assessment was optional, only slightly more than 50% returned for echocardiographic assessment at 1 year. Furthermore, only rudimentary baseline information was available (e.g. preimplant pacemaker rate was not assessed). Baseline parameters would have been important to gain a better understanding of how the population changed between SAVI-1 and SAVI-2. Lastly, future research should include long-term follow-up.