All patients older than 18 years old with severe symptomatic aortic stenosis, who underwent TAVI between January 2016 and December 2018 in a tertiary hospital, were analyzed retrospectively. Patients with pre-existing PPM, nontransfemoral approach, valve-in-valve procedures and intraprocedural death were excluded.

The local ethics review board approved this study and written informed consent was waived due to the study's retrospective nature. The study was conducted following the declaration of Helsinki.

All patients proposed for TAVI were evaluated by a multidisciplinary heart team and the eligibility for TAVI procedure was decided by consensus. Pre-procedure cardiac and iliofemoral access computed tomography angiography were performed in all patients. The type of prosthesis (self-expandable vs balloon-expandable) was chosen by the heart team.

Overall, patients were electively admitted to the hospital on the day of the procedure, and all underwent transfemoral TAVI in a cardiac catheterization laboratory. The procedures were performed under local anaesthesia and minimal sedation, with continuous haemodynamic and electrocardiographic monitoring. Femoral artery access was obtained for all patients, under fluoroscopy or ultrasound guidance. Femoral punctures were routinely preclosed using 2 Perclose ProGlide devices (Abbott Vascular, United States) or, more recently, one Perclose and one Angio-Seal 8-Fr (Terumo, Japan).

Pacing method was at the operator's discretion. In most cases, the temporary pacemaker was removed in the catheterization laboratory if considered suitable by the operator.

After the procedure, all patients were admitted at the cardiac intensive care unit.

The indication for PPM implantation and timing of discharge were at the discretion of the treating physician and in accordance with the pacing and electrophysiology laboratory team. No electrophysiological studies were conducted.

Clinical, procedural, electrocardiographic and follow-up data were collected through hospital records. Patients were followed by regular outpatient visits at our centre or the referring hospital; decisions regarding the frequency of visits were at the treating physician's discretion.

Variables definitions are presented in supplementary data.

Standard 12-lead electrocardiograms (ECG) were collected at baseline, immediately after the procedure, daily until day 3 post-TAVI and at discharge. Electrocardiograms were recorded using standard calibration (10mm=1mV, 25mm/s). The ECGs were retrospectively analyzed by the same investigator and the following parameters were assessed: rhythm, PR interval, presence, and type of atrioventricular (AV) block, QRS interval, presence of bundle-branch block, left anterior fascicular block, left posterior fascicular block or nonspecific intraventricular conduction disturbance with QRS interval≥120ms. All these analyses were performed according to the recommendations for the standardization and interpretation of ECG,16 and based on the definitions of CD after TAVI, proposed by the 2019 consensus document15 (supplementary data). Additionally, continuous telemetry monitoring was recorded during the procedure and during all hospital stay.

According to the 2019 Scientific Expert Panel consensus document,15 and after analysing the information obtained from the baseline 12-lead ECG, continuous telemetry monitoring during the procedure and 12-lead ECG immediately after the procedure, patients were retrospectively classified in 5 groups (Fig. 1): group 1, no ECG changes in patients without pre-procedure right bundle brunch block (RBBB); group 2, no ECG changes in patients with preexisting RBBB; group 3, ECG changes in patients with preexisting CD; group 4, new-onset left bundle brunch block (LBBB) and group 5, HAVB or complete heart block (CHB) during the periprocedural period.

Fig. 1.

Group allocation and criteria for early discharge according to the conduction disturbances following TAVI. Adapted algorithm from the 2019 Management of Conduction Disturbances Associated With Transcatheter Aortic Valve Replacement. Modified with permission from Rodés Cabau et al.15 AVB, atrioventricular block; CHB, complete heart block; ECG, electrocardiogram; HAVB, high-degree atrioventricular block; IVCD, intraventricular conduction delay; LBBB, left bundle brunch block; RBBB, right bundle brunch block. *1st degree AVB, RBBB, LBBB or QRS≥120ms with IVCD.

(0.74MB).

The proposed algorithm15 (Fig. 1) was adapted and retrospectively applied in our cohort to identify which patients could have been eligible for ED with no further measures, and which patients were not suitable for ED, due to a higher risk for HAVB/CHB or because they were candidates for PPM implantation (for more details regarding ED criteria, refer to supplementary data).

Early-discharge was defined as discharge at day-3 after the procedure.

The primary endpoint was defined as the rate of PPM implantation within 30 days after TAVI regarding eligibility for ED. Secondary endpoints included 6-month atrial and/or ventricular pacing rates, 1-year rate of syncope and all-cause mortality and the composite outcome of both, in relation with ED eligibility.

Descriptive statistics for categorical variables were summarized as absolute and relative frequencies and continuous variables were expressed as median and interquartile range [IQR].

Patients’ characteristics were compared between patients with and without ED using χ2 test or Fisher's exact test, as appropriate; or Student t test or Mann–Whitney U test, in case of non-Gaussian distribution using the Kolmogorov–Smirnov test.

Statistical significance was set at a 2-tailed probability level of <0.05. All statistical analysis was performed using Statistical Package for Social Sciences (SPSS) software version 23.0 (IBM, Armonk, USA).

A total of 220 patients underwent transfemoral TAVI between January 2016 and December 2018. The median age was 82 years [76–86] and 126 patients (57%) were women. Median Society of Thoracic Surgery mortality score was 4.356% [2.861–6.275] and a self-expandable prosthesis was used in 146 patients (66%). Baseline demographic, clinical and procedural characteristics of the study sample are summarized in Table 1.

Patients’ allocation to the 5 pre-defined groups is shown in Fig. 2. Seventy-eight patients (35%) were categorized in group 1, 11 patients (5.0%) in group 2, 13 patients (5.9%) in group 3, 80 patients (36%) in group 4 and 38 patients (17%) in group 5.

Fig. 2.

Group allocation in the overall population and regarding early-discharge eligibility. ED, early-discharge.

(0.24MB).

The 30-day PPM implantation rate as a function of group allocation is shown in Fig. 3. Group 5 showed the highest PPM rates (70.3%), followed by group 3 (30.8%).

Fig. 3.

30-Day permanent pacemaker implantation rate after transcatheter aortic valve implantation in the overall population, according to group allocation. PPM, permanent pacemaker implantation; TAVI, transcatheter aortic valve implantation.

(0.07MB).

One hundred and forty-five patients (65.9%) were considered ED candidates and 75 patients (34.1%) non-ED candidates. Baseline demographic, clinical and procedural characteristics comparing ED with non-ED candidates are present in Table 2. Early-discharge candidates presented a significant higher median left ventricle ejection fraction (55% [IQR 53–59] vs 53% [IQR 38–57]; P=.003) and lower median prosthesis size (26mm [23–29] vs 26mm [25–29]; P=.029) than non-ED candidates. There were no significant differences between ED and non-ED candidates regarding the type of valve implanted (self-expandable vs balloon-expandable), although there was a greater tendency for higher rates of ED in patients who underwent balloon-expandable valve implantation (74% vs 62%; P=.061) (Fig. 1 of the supplementary data).

Baseline median PR (172ms [151–200] vs 180ms [152–223]; P=.14) and QRS interval (102ms [90–118] vs 105ms [94–142]; P=.14) were similar between ED and non-ED candidates, as well as the prevalence of pre-existing LBBB (12% in both groups; P>.9). However, ED candidates had a significantly lower prevalence of RBBB prior intervention (7.6% vs 27%; P<.001).

The association between ED eligibility and group assignment according to the consensus document is shown in Fig. 2. Early-discharge candidates were classified in group 1 significantly more often than non-ED candidates (45% vs 17%) and less often in group 5 (9.0% vs 33%) (overall P value<.001).

The overall median length of hospital stay until discharge was 7 days [IQR 6–8]. Only 7 patients were effectively discharged within 3 days after TAVI. Of those, 3 patients were prospectively classified as ED candidates and 4 patients as non-ED candidates. The latter underwent PPM implantation during the first days of hospital stay, allowing discharge within 3 days post-TAVI. There were no significant differences in median length of hospital stay between patients classified as ED (7 days [IQR 5–8]) and non-ED candidates (7 days [6–9.5]); P=.116.

The 30-day rate of PPM implantation for ED candidates was 3.6% and for non-ED candidates was 62.5% (P<.001), all implanted during index hospital stay. The median time until PPM implantation for ED candidates was 6.0 days [5.0–9.5] and 2.0 days [1.0–5.0] for non-ED candidates (P=.011).

Five patients eligible for ED underwent PPM implantation within 30 days after TAVI: one patient was assigned to each of groups 1, 3 and 4, and 2 patients to group 5.

The majority of these patients (3 patients) underwent PPM implantation due to sinus node dysfunction. Of those, none presented syncope or haemodynamic instability. The 6-month rate of atrial and ventricular pacing was: a) 47.5% and 11.3%, b) 88% and 93%, and c) 89.6% and <1%, respectively. The remaining 2 patients underwent PPM implantation due to CHB (at day 5 and 6 after TAVI); one presented syncope and required temporary transvenous pacemaker. The 6-month rate of ventricular pacing was 10% and 2%, respectively.

None of these patients showed pacemaker dependency at 6-month follow-up.

The 6-month rate of atrial pacing was higher in ED candidates (47.5% [1.0–88.8] vs 14.5% [3.4–39.5]; P=.546) and the rate of ventricular pacing was lower in ED candidates (10.0% [1.0–93.0] vs 85.5% [8.5–99.0]; P=.146), although the difference was not statistically significant (P>.05). No differences were reported regarding 1-year mortality (4.9% vs 6.9%; P=.750), 1-year prevalence of syncope (2.9% vs 4.7%; P=.820) and the composite outcome of 1-year mortality or syncope (7.9% vs 11.8%; P=.510), between ED and non-ED candidates.

Seventy-five patients (34.1%) were not eligible for ED: 35 patients (15.9%) were classified as candidates for PPM implantation and 40 patients (18.2%) as at higher risk for HAVB/CHB.

Candidates for PPM implantation were more often assigned to group 5 (68.6% vs 2.5%) and less to group 4 (11.4% vs 57.5%) than patients at higher risk for HAVB/CHB (Table 3).

According to the suggested algorithm, 97.1% of the patients classified as candidates for PPM implantation and 29.7% of the patients classified as at higher risk for HAVB/CHB effectively underwent PPM implantation within 30 days after TAVI (Fig. 4). PPM was implanted on average at day one and six after TAVI, for PPM implantation candidates and for patients at higher risk for HAVB/CHB, respectively. The 6-month median atrial and ventricular pacing rates was 9.9% [1.4–21.5] and 97.5% [53.3–99.0] for PPM implantation candidates and 37% [22.3–76.4] and 3.4% [1.0–36.1] for patients at higher risk for HAVB/CHB, respectively.

Fig. 4.

30-Day permanent pacemaker implantation rate after transcatheter aortic valve implantation, according to the algorithm recommendations of the 2019 consensus document. CHB, complete heart block; ED, early-discharge; HAVB, high-degree atrioventricular block; PPM, permanent pacemaker; TAVI, transcatheter aortic valve implantation.

(0.08MB).

There were no significant differences between the type of valve implanted (self-expandable vs balloon-expandable) regarding group allocation and according to the algorithm recommendations of the 2019 consensus document (Figs. 1–3 of the supplementary data). However, patients with balloon-expandable valves were more frequently allocated to Group 1 (46% vs 30%) and less frequently to Group 4 (30% vs 40%) and there was a lower tendency for them to be classified as higher risk for HAVB/CHB (10% vs 23%; P=.053). There were no significant differences in the 30-day rate of PPM implantation between the type of valve implanted (P=.296).

The high prevalence of CD after TAVI, along with the high rate of PPM implantation, and the uncertainty regarding the timing of discharge, render the management of patients who develop CD post-TAVI a challenge. A recent multidisciplinary expert consensus document has been published to provide a standardize pathway for the management of CD after TAVI.15

In our cohort, the adapted algorithm identified a significant number of patients (66%) as eligible for discharge within 3 days after TAVI. In those patients, the prevalence of pre-existing LBBB and RBBB was 12% and 7.6%, respectively. Almost half of the ED candidates (45%) were classified in group 1 (no ECG changes in patients without RBBB pre-procedure). Malebranche et al.17 validated the same algorithm in a large cohort and also found that most of the patients (73%) were eligible for ED, with similar prevalence of LBBB and RBBB as in our study (11% and 6%, respectively). However, the incidence of patients assigned to group 1 was higher than in our cohort (59% vs 45%), which may explain the increased prevalence of patients suitable for ED than in our study.

We found that patients eligible for ED had a lower prevalence of pre-existing RBBB than non-ED candidates (7.6% vs 27%), without differences in the prevalence of baseline LBBB.

Although in the proposed algorithm, baseline RBBB doesn’t preclude an ED strategy, the clinical management of these patients is still a matter of great debate. Pre-existing RBBB seems to be the strongest and most consistent risk factor for PPM implantation after TAVI.11,18 Auffret et al.19 found that patients with baseline RBBB who underwent TAVI and were discharged without PPM during the index hospitalization, were at higher risk of HAVB and/or sudden cardiac death during follow-up. Another study found that all-cause mortality and cardiovascular mortality was higher in patients with pre-existing RBBB.20 The higher incidence of CD in patients with baseline RBBB make them less suitable for an ED strategy, which is consistent with our findings. However, in our cohort, non-ED candidates didn’t show an increase in 1-year mortality, despite having a higher prevalence of baseline RBBB. This could be explained because almost 65% of non-ED patients underwent PPM implantation, protecting them from HAVB and making them less vulnerable to sudden cardiac death.

Despite no significant differences were found regarding eligibility for ED between the type of valve implanted (self-expandable vs balloon-expandable), there was a greater tendency for higher rates of ED in patients who underwent balloon-expandable valve implantation. Patients with balloon-expandable valves were more frequently allocated to Group 1 and less frequently to Group 4. Additionally, there was a lower tendency for them to be classified as higher risk for HAVB/CHB. It is well established that self-expandable valve implantation is a major predictor of CD after TAVI,21 which is in line with the findings of our study. These differences may explain the higher number of patients with balloon-expandable valves classified as suitable for ED in our cohort.

However, due to the relatively small sample size and because it is a secondary analysis in a subgroup, this may have contributed to not finding significant differences between the type of valve implanted.

The 30-day rate of PPM implantation for ED candidates was 3.6%, similar to the rate reported by Malebranche et al.17 (2.7%). In our cohort, ED candidates underwent PPM implantation on average 6 days after TAVI and the majority had a PPM due to bradycardia–tachycardia syndrome, to enable the initiation and titration of pharmacological treatment for atrial fibrillation and to protect from the prolonged sinus pause following the termination of tachycardia, which translated into high rates of atrial pacing during follow-up. In general, pacing for sinus node dysfunction does not seem to affect survival,22 as opposed to pacing for HAVB/CHB.

Also, no significant differences were observed regarding 1-year mortality, syncope, or the composite outcome of both, between patients eligible for ED and those who were not.

When applied to the real-world practice, and in line with the latest evidence,17 the proposed algorithm appears to identify low-risk patients, who could have been ED. Also, it presented a high predictive negative value for subsequent need for PPM implantation, allowing an accurate recognition of patients suitable for ED and that will not benefit from PPM implantation.

However, since the results of this study are based on a retrospective analysis and as we did not discharged patients who would be candidates for ED, we may not have detected all potential adverse events related to CD if the patient had been discharged.

This is a retrospective analysis, subject to the limitations of this type of study. Patients were managed at the physician's discretion and not as proposed by the algorithm which was implemented later and did not interfere with the medical decision. The indication for PPM implantation was at the discretion of the treating physicians, and therefore different criteria may have been used for decision about PPM implantation among patients subsequently classified as eligible or non-eligible for ED, introducing potential bias and limiting the validity of the findings. A prospective study design with a standardized protocol for PPM implantation would provide more robust evidence.

This study has a relatively small sample size, which could limit the generalizability of the findings, especially in multiple subgroups. Despite this, the main findings of our study were in agreement within a larger cohort from the study of Malebranch et al.17

Finally, electrophysiological study was not performed as well as systematic electrocardiographic monitoring after discharge.