Discussion
Valve-in-valve TAVR is a seductive option in case of degenerated bioprostheses. Apparently the advantages are obvious: the procedure is safe, easy and convenient to perform, positioning of the valve is simplified due to radiopaque markers and paravalvular leakages as a risk factor for mortality are not an relevant issue [1-3, 12, 13]. Furthermore, waiving surgery is a striking and convenient point for the patient. Therefore, the crucial point remains durability. If the initially unargued excellent results proof to be stable over time, VIV-TAVR is ready for the many. If not, it must remain a bailout-option for risky patients. The presently available literature does not allow to answer this crucial question conclusively [13-18].
The primary clinical efficacy of VIV-TAVR is unargued. Regarding this point, the present study continuous excellent early results of our group and is in line with the data given by the VIVID-registry [3, 13]. The present patient cohort represents a typical TAVR patient population. Likewise, procedural data and hospital outcome are comparable with VIVID [13]. The procedure itself is convenient and technically less demanding - radiopaque markers of the bioprosthesis clearly indicate the perfect landing zone and facilitate an easy and orthograde positioning of the valve. Additionally, the sealing is much more effective compared to “native”-TAVR and though, paravalvular leakages are no matter of greater concern in VIV-TAVR as the present study and the VIVID-registry could demonstrate [20]. Nonetheless, one point has to be taken into account: coronary obstruction. Despite rare with an overall incidence of 0.7%, coronary obstruction is an issue being associated with a high mortality up to 42% [21]. To prevent from this fatal course, some considerations should be taken into account during procedure planning. There exist some predicting variables like female gender, low coronary ostial height (<10mm), Sinus Valsalva width <30mm and presence of biological valves with externally mounted leaflets or even worse, stentless valves [22]. For those conditions the BASILICA procedure or alternatively the OPEN-BASILICA-procedure were described before [22-24]. Anyhow, in the present series coronary obstruction played no role.
Generally, the procedural and hospital outcomes observed in this study confirmed the safety of VIV-TAVR. Indeed, the hospital survival as well as stroke rate were better than the results reported in the VIVID (1.3% vs. 7.6% for mortality and 0.0% vs. 1.7% for major stroke).
During the implantation, particularly with self-expanding prostheses it is aimed to release the valve in a pronounced high position to ensure supra-annular positioning and to allow a good hemodynamic result. According to this supra-annular concept, the immediate hemodynamic outcomes are superior with self-expandable prostheses compared to intra-annularly implanted balloon-expandable valves. This fact could be confirmed by the present study demonstrating a significantly higher rate of early device success according to the VARC-2 criteria for balloon-expandable devices. These observations made are mainly in line with those reported by the Dvir and colleagues [13, 18].
Concluding procedural aspects and initial outcomes, it seems to be evident, that VIV-TAVR as a heterogeneous group consisting of different procedures (by means of combinations of surgical and TAVR valves) - with respect to some special cases - basically is safe and provides good or at least acceptable initial hemodynamic results [3, 13, 16]. Furthermore, it is evident, that VIV-TAVR in small surgical valves as well as stented valves is associated with higher postprocedural gradients [16].
As far as good – but Dvir and colleagues identified a sore spot already in the VIVID [13]. Patients presenting with an internal diameter equal or less 20mm of the surgical valve showed to have an inferior 1-year survival [13]. Accordingly, the PARTNER 2 Valve-in-Valve Study at 3-years priorly defined a labeled size <21mm as a key exclusion criterion [17]. The present study confirms these findings. Despite experiencing an initial sufficient reduction of transvalvular pressure gradients in true-ID’s equal or less 21mm, gradients continuously were significantly higher up to 1-year of follow-up, compared to larger valve. After 2 and 3-years the gradients stayed to be high, but not anymore being significantly higher compared to larger valves. This development could be explained by two facts. First, the larger valves likewise experienced a fundamental increase in transvalvular pressure gradients and second, a potential bias within the small-ID group due to patients with assumably higher gradients dying earlier. This explanation fits into the observed higher mortality of the small-ID group during follow-up. Bleiziffer et al. reported in the 5-years outcome of the VIVID likewise a significantly increased mortality in small-ID patients [18]. Hence, the presently available results allow to conclude, that VIV-TAVR in patients with a small ID (≤21mm or ≤20mm depending on the definition) are likely to have inferior outcomes after their primary hospital stay.
A quite more unclear situation applies to larger valves with an ID ≥ 21mm. For example, Webb and colleagues reported in the 3-Year outcomes of 365 patient in the PARTNER 2-registry sustained hemodynamic status with at least only minimal changes in mean transvalvular pressure gradients [17]. The observation we made in the present series are contradictory: the predischarge hemodynamic outcomes were quite comparable to those reported by Webb and colleagues (16.8 ± 7.1 mmHg vs. 17.4 mmHg) [17] - despite 100% use of balloon-expandable valves in the PARTNER 2-registry and the large portion (68.8%) of self-expanding devices in the present series.
In the present series the mean gradients steadily increased over the following 3-years (Figure 2A ), which happened independently from the true-ID of the surgical valve (Figure 2B ) or the type of implanted transcatheter valve (Figure 3 ). This could not be explained by size of the surgical valves: the ratio of patients with a true-ID less or equal 21mm was even higher in the PARTNER 2-registry compared to our data (76.7% vs. 58.4%).
Up to date, there exist no further long-term hemodynamic data in the present literature for further comparison. Most studies dealing with hemodynamic results, end after 1-year of outpatient follow-up [14-16]. Accordingly, hemodynamic long-term-follow-up has much more to be elaborated in further studies.
A particular matter of concern beside hemodynamics is survival. All presently available essential long-term-trials describe mainly comparable survival outcomes [17, 18, 25, 26]. In the PARTNER 2-registry Webb and colleagues reported an estimated all-cause mortality of 32.7% after 3-years of follow-up, which fits to 27.7% mortality after 3-years described in the CoreValve US Expanded Use Study [17, 25]. The data given by the present study is with an estimated mortality of 42.9% after 3-years mainly in range, but being somewhat higher.
Further long-term-outcomes are provided by a small multicenter study with 116 patients reporting 5-years mortality of 32.1% as well as by the long-term data of the VIVID with 62.0% mortality after 8-years [18, 26]. In the present series a poor estimated 7-years survival of 20.6 ± 16.1% was found. Interpreting these poor survival data should respect the average age ranging between 76.0 ± 11.0 and 79.4 ± 5.8 years [17, 18, 25, 26].
Meanwhile the large VIVID-registry identified smaller true-IDs as a risk factor for mortality beginning from 1-year of follow-up, the smaller PARTNER 2-registry as well as the present study could not confirm that observation findings [13, 17, 18]. Potentially, the overall high mortality rate in the study population conceals possible significant differences.