Transcatheter aortic valve implantation with a mechanically expandable prosthesis: a learning experience for permanent pacemaker implantation rate reduction

Patients undergoing trans-femoral TAVI with new generation LOTUS prosthesis were included. All patients in this series had been diagnosed with severe symptomatic A-V stenosis. Patients with the following conditions were excluded from the present analysis: pure aortic insufficiency, previous-AV surgery (replacement/repair), previous TAVI, and previous PPMI. Moreover, we have excluded from this series two patients where TAVI with the LOTUS was abandoned in favor of TAVI with a prosthesis of different model and larger size (one patient with 27.8 mm aortic annulus) or to convert to conventional aortic valve replacement (one patient developing left ventricular perforation during TAVI with LOTUS). Patients included in the present series were treated in the period 2014–2016 by our TAVI team.

Perioperative data were prospectively collected and reported according to the valve academic research consortium definitions (VARC).

All procedures were performed under fluoroscopic and trans-esophageal monitoring. Vascular access was percutaneous and through the right femoral artery in the majority of patients. Since the beginning of our experience, we have not performed aortic balloon valvuloplasty when implanting the LOTUS valve. Prosthesis selection has been routinely based upon preoperative evaluation of cardiac computed tomography images reconstructed using both the OSIRIX (Pixmeo SARL, Bernex, CH) and 3mensio (Pie Medical Imaging, Maastricht, NL) software. Since the beginning of our experience with the Lotus valve, we have adopted a policy of maximum 10% oversizing of the prosthesis in reference to the native annulus area-derived diameter.

After probing the native aortic valve, a pre-shaped stiff guidewire was used to advance the Lotus prosthesis. Details concerning the design and implantation of the Lotus prostheses have been previously described in the literature [2]. No patient received the most recent version of the Lotus prosthesis (i.e., Lotus Edge).

At the end of each procedure, prosthesis implantation depth was measured at the level of the non-coronary and left coronary cusps starting from the control aortography, in the nadir view. Measurements were obtained off-line after the procedure by an investigator that did not take part in the valve implantation.

Through our experience with the LOTUS valve, we have designed a structured protocol to reduce PPMI rate. After being reviewed by our internal research/ethics committee, and being submitted to the Berlin Medical Council, the protocol was implemented in the second phase of our experience with the Lotus (group B) and it is nowadays part of our standard operating procedure for TAVI.

Data are presented as absolute numbers, percentages, mean ± standard deviation for normally distributed variables and median with minimum and maximum values for variables with not-normal distribution. Patients that did and did not require PPMI after TAVI as well as patients being treated before (group A) and after the introduction of our PPMI reduction protocol (group B) were compared. Differences of preoperative clinical/anatomical variables and perioperative management (including the PPMI reduction protocol) were analyzed. Paired Student t test, Wilcoxon signed-rank test, χ², and Fischer-exact test were used whenever appropriate. A p-value < 0.05 was considered as significant.

All surviving patients underwent at least one follow-up visit, including surface ECG analysis, within 1 month from discharge.

No new PPMI was necessary. Out of the 14 patients implanted with PPM, nine (64%) were PM-dependent at follow-up (eight in group A and one in group B).

At a median clinical follow-up of 209 days (30–721 days), estimated survival in patients with PPMI after TAVI was 85.7 and 88.7% in those that did not require PPMI (p = 0.3) (Fig. 2).

None of the surviving patients required a new PPMI and none developed complications resulting from the perioperative management.

The LOTUS TAVI prosthesis is mechanically expandable and presents some peculiar features. Due to its elevated radial force, this prosthesis results in high rates of full expansion. At post-procedural computed tomography evaluation, the LOTUS prosthesis shows almost complete circularization of the native annulus that results in low trans-prosthetic gradients and minimal paravalvular leak rates [9]. Its particular radial strength, exerted within the native annulus, may also lead to an increased compression on the bordering anatomical structures, such as the cardiac conduction tissues. It has to be emphasized that the very last design of the Lotus valve (Lotus Edge), that presents an improved profile of catheter and prosthesis to diminish left ventricular outflow tract traumatism, was not employed in the presented series.

Since the beginning of our experience with the Lotus prosthesis, we have adopted a concept of minimal prosthesis oversizing, to reduce the trauma exerted on the landing zone. As a result, the prosthesis/annulus size ratio is similar in the two temporal cohorts. As documented in our results, PPMI may be necessary even in patients treated with minimal valve oversizing. This is possibly due to the prosthesis implantation technique and to the final position of the valve, at time of release. In fact, the relationship between implantation depth and PPMI rate has been shown since the beginning of the TAVI experience with the Medtronic Core-Valve System (Medtronic Inc., Minneapolis, Minnesota) [10] and it has been restated in more recent studies [11]. Very recently, Zaman et al. have confirmed in a series of 95 Lotus patients, a PPMI rate of 28% [12]. When chasing for independent predictors of PPMI, pre-existing RBBB [hazard ratio (HR) 2.8, 95% CI 1.1–7.0; p  =  0.032] and depth of implantation below the non-coronary cusp (HR 2.4, 95% CI 1.0–5.7; p  =  0.045) were the strongest determinants [12].

It is reasonable to think that, apart from the final position of the valve within the annulus, the intermediate phases before achieving final valve release may also impact upon PPMI rate. In fact, trauma to the conduction tissue can happen at any time during the valve release process and result from a direct injury exerted by the valve frame and/or valve catheter. This is of particular importance when adopting a mechanically expandable prosthesis with elevated radial force such as the LOTUS prosthesis. In light of these considerations, the core of our proposed LOTUS implantation strategy is to minimize, at any given point during the procedure, impingement of the prosthesis and catheter too deep into the left ventricular outflow tract. To reduce this eventuality, we have tried to monitor the valve depth within the aortic annulus during the various phases going from valve advancement into the left ventricular outflow tract, valve unsheathing, complete expansion, and shortening. As already emphasized in the methods section, we propose an implantation technique where the nadir of the valve metallic frame should be kept, at any given time during the procedure, immediately below the annular plane (approximately 5–10 mm) using a pigtail as reference in the deepest point of the non-coronary sinus. To achieve this goal, we have also observed an increment in the valve re-sheathing rate. The re-sheathing was actually adopted to fine-tune and optimize the final position of the valve. Our implantation differs from that originally proposed by the manufacturing company and actually routinely applied in the first phase of our experience with the LOTUS. In fact, the manufacturer suggests implanting the valve using as reference point the radiopaque marker located in the middle of the valve nitinol frame. By following the manufacturer indication the prosthesis lays, during the various phases of valve positioning, deeper within the annulus and LVOT with at least half of its frame below the nadir of the native annulus. While the valve shortens and reaches its final position, it squeezes and pulls upward whatever tissue it has anchored during the earlier phases of positioning.

When looking more specifically at the prosthesis final position, it is now accepted that the shallower the implantation depth the lower the PPMI rate. Petronio et al. have confirmed, by means of receiver operating curve analysis, that a Core-valve implantation depth of < 4 mm at the level of the non-coronary cusp is a negative predictor of PPMI [11]. In fact, the bundle of His passes through the right fibrous trigone and the membranous ventricular septum and gives off fibers that form the left bundle beneath the commissure between the right and non-coronary cusps, in close vicinity to the aortic valve, over a distance of 6.5–20 mm [13].

The negative predictive values of PPMI according to prosthesis depth resulting from both ours and Petronio et al. analysis [11] support the fact that, when implanting the prosthesis shallower than a certain depth cut-off, the chances of not requiring a PPMI are around 90%. On the contrary, once the depth cut-off has been overcome and the valve is implanted deeper within the annulus, the probability of requiring a pacemaker is around 20% (positive predictive values of, respectively, 24.8 and 21.7%). This confirms that, apart from prosthesis final depth, there are many other clinical and perioperative factors that may impact upon PPMI rate.

In synthesis, our results substantiate the importance of implantation depth even when using the LOTUS device [12] and we were actually able to identify an ideal cut-off after which the risk of PPMI should increase. It should be noticed that in our analysis depth at the level of the left coronary sinus, instead of the non-coronary sinus, seemed to play a stronger and independent impact upon PPMI rate. In fact in our experience, in spite of the self-centering properties of the LOTUS prosthesis, once the valve is fully released the nadir of the nitinol frame is most often at the level of the left coronary cusp.

We have structured part of our proposed PPMI reduction protocol, including dexamethasone and orciprenaline treatment, on the basis of previous findings in the field of percutaneous treatment of congenital heart disease. After percutaneous closure of perimembranous ventricular septal defects, cardiac conduction delays may result directly from mechanical compression, inflammation, and fibrosis. The septal occluder device has been shown to cause a localized inflammatory reaction that can result in extensive scar tissue and cartilaginous metaplasia of the surrounding myocardium [14]. Conduction abnormalities after TAVI may have a similar ground and the proposed protocol, although completely original in the TAVI field, has its own logic.

The use of steroids has been initially reported to be successful for treatment of atrio-ventricular block in inflammatory diseases [15]. More recently, at least two groups have proposed an empirical regimen of steroids, non-steroidal anti-inflammatories, and chronotropic agents to manage iatrogenic heart block after percutaneous closure of ventricular septal defects [16, 17].

Although our initial findings seem encouraging (with one patient experiencing regression of an intermittent complete A-V block and three patients showing regression of an initial LBBB > 130 ms), we do not have enough data to conclude that the PPMI rate was actually reduced thanks to the application of this pharmacological regimen. In fact, acquisition of confidence and dexterity with the Lotus valve was possibly the main factor that allowed us to reduce the PPMI rate as we gained experience. Moreover, we have to remark that in three cases complete A-V block did not improve after 5 days of treatment with positive chronotropic agents (in three patients) and steroids (in only two patients). Finally, the risks of high-dose steroidal treatment should be considered. In our experience, dexamethasone administration did not lead to major morbidity and/or mortality and we limited its use to patients developing complete A-V block and LBBB with widening QRS (> 120 ms). In any case at the present, we are testing the pharmacological protocol in a larger sample, including patients implanted with other valve types.

In the past, prophylactic PPMI has been proposed, and used, also in patients with new onset of LBBB and prolonged PR interval after TAVI, not in line with the international recommendations for PPMI [3]. The possibility of reducing PPMI rate after TAVI by strictly adhering to international guidelines for PPMI has been just recently proposed [3, 11, 18].

We have categorically started to follow the international guidelines for PPMI just in the second phase of our experience with the LOTUS valve (PPMI reduction protocol in group B). Using this more conservative approach, no total A-V block or sudden death occurred between discharge and follow-up, even in patients that had developed LBBB after TAVI. This is an important point because it is not yet clear how a widening QRS complex is going to evolve after implantation of a TAVI prosthesis with such high radial force.

Timing for PPMI should also be viewed critically to minimize PPMI rate. In both the REPRISE II trial and RESPOND post-market study, the majority of PPMs were implanted within 2 days from TAVI with LOTUS, and more than half within 24 h. Interestingly, less than 50% of patients undergoing PPMI were pacemaker-dependent after 30 days [2, 8].

These findings confirm the fact that injuries to the conduction system occurring during TAVI may be temporary, mainly related to localized edema and inflammation and consequently may heal in due time. A closer look to the surgical literature shows that in some centers it is policy waiting until at least the 5th postoperative day before performing PPMI after conventional aortic valve replacement [19]. To safely optimize the timing of PPMI after TAVI, we have revised our management of temporary pacemakers. Introducing the temporary PM via the jugular vein instead of the femoral vein may be advisable to reduce the infection risk and to allow for patient mobilization, whenever the PM is kept in place to wait for recovery of normal cardiac conduction. Although keeping a temporary lead in place for prolonged time may generate concerns, a more liberal approach towards PPMI will also add invasiveness and iatrogenic risks. In any case in our experience, a temporary PM was kept in place for up to 5 days after TAVI in six patients. Patients were kept on telemetry in a high-dependency unit and in none of them the temporary lead resulted in iatrogenic complications such as cardiac perforation and/or infection.