Impact of Bundle Branch Block on Permanent Pacemaker Implantation after Transcatheter Aortic Valve Implantation: A Meta-Analysis

Data regarding the impact of infra-Hisian conduction disturbances leading to permanent pacemaker implantation (PPI) after transcatheter aortic valve implantation (TAVI) remain limited. The aim of this study was to determine the impact of right and/or left bundle branch block (RBBB/LBBB) on post-TAVI PPI. We performed a systematic literature review to identify studies reporting on RBBB and/or LBBB status and post-TAVI PPI. Study design, patient characteristics, and the presence of branch block were analyzed. Odds ratios (ORs) with 95% CI were extracted. The final analysis included 36 studies, reporting about 55,851 patients. Data on LBBB were extracted from 33 studies. Among 51,026 patients included, 5503 showed pre-implant LBBB (11.9% (10.4%–13.8%)). The influence of LBBB on post-TAVI PPI was not significant OR 1.1474 (0.9025; 1.4588), p = 0.2618. Data on RBBB were extracted from 28 studies. Among 46,663 patients included, 31,603 showed pre-implant RBBB (9.2% (7.3%–11.6%)). The influence of RBBB on post-TAVI PPI was significant OR 4.8581 (4.1571; 5.6775), p < 0.0001. From this meta-analysis, the presence of RBBB increased the risk for post-TAVI PPI, independent of age or LVEF, while this finding was not confirmed for patients experimenting with LBBB. This result emphasizes the need for pre-operative evaluation strategies in patient selection for TAVI.


Introduction
Atrio-ventricular conduction disturbances and subsequent permanent pacemaker implantation (PPI) represent frequent complication after transcatheter aortic valve implantation (TAVI) [1,2]. Notably, infra-Hisian conductions' desynchrony such as left bundle branch block (LBBB) and right bundle branch block (RBBB) remain an ongoing issue in TAVI [3,4], especially considering TAVI as well-established therapeutic approach for patients with aortic stenosis at high surgical risk [5], while considerable advances in procedural techniques tend to extend TAVI indications to patients with a lower surgical risk [6].
Mechanical stress on the aortic valve annulus, deterioration of the ventricular septum, and local edema may all injury the atrio-ventricular conduction during the TAVI procedures [1,3]. In patients with pre-existing conduction system impairments, such additional procedural-related factors may contribute to a higher post-TAVI PPI rate. Consequently, complications such as PPI thus remain a substantial barrier to extending this technique to operable patients who would otherwise undergo surgery [7]. Therefore, better patient selection and identification of pre-operative risk factors for progression of conduction disturbances, and subsequently PPI, are decisive [8]. Current data about the clinical impact of bundle branch block on post-TAVI PPI remain controversial [9,10]. Left bundle branch block (LBBB) occurs in 5 to 65% in TAVI patients and leads to PPI in 15 of 20% of cases [11]. Pre-operative right bundle branch block (RBBB) is present in 10 to 21% of patients and results in up to 40% of post-operative PPI, making pre-operative RBBB the most important patient-related factor [2,12]. However, the prognostic value on pre-existing infra-Hisian conduction disturbances on post-TAVI PPI remains unclear [10][11][12].
We aim to investigate the clinical impact of pre-operative RBBB/LBBB on PPI after TAVI.

Research Strategy
This meta-analysis was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and the research strategy was developed according to available guidance from the Cochrane Collaboration. A broad, computerized literature search was performed to identify all relevant studies from Embase, Cochrane database, and PubMed exploring Medical Subject Heading (MeSH) terms related to pre-operative RBBB or pre-operative/new-onset LBBB in TAVI population. The PubMed database was searched entering the following key words: "Pacemaker, Artificial" [Mesh] OR pacemaker implantation AND "Transcatheter Aortic Valve Replacement" [Mesh] OR transcatheter aortic valve replacement AND "Bundle-Branch Block" [Mesh]. We restricted the research to English publications. Last access to the database was on 1 November 2020. The search was limited to studies in human.

Eligibility Criteria and Studies Selection
Studies were included in the final analyses if patients were >18 years (I); >250 patients were included in the main analysis, in order to provide data interpretation of the most consistent clinical series (II); and studies provided a description of pacemaker status of the population (III). Furthermore, articles with no possible extraction of the presence of RBBB/LBBB were excluded. Pre-operative RBBB/LBBB and new-onset RBBB/LBBB were both included in the present analysis. Systematic review and meta-analyses were not taken in account. Studies describing cardiac surgery procedures were also left out. The selected articles underwent extensive evaluation at title and abstract level by two independent researchers (J.R. and M.D.M.) to assess the potential inclusion in the metaanalysis. Discrepancies were solved by consensus with the intervention of a third reviewer (R.L.). There were no duplicate data.

Statistical Analysis
Calculation of an overall proportion from studies reporting a single proportion was performed using a meta-analytic approach by means of metaprop function of meta package in R. A logit-transformation was performed as suggested by Warton & Hui [13] to calculate confidence intervals (CIs) for individual study results. A Clopper-Pearson approach and a DerSimonian-Laird estimator were used to estimate the between-study variance [14]. Total proportion with 95% Cl was reported. Funnel plot and Egger's test were used for estimation of publication bias. The primary endpoint was 30-day or in-hospital PPI after TAVI, so odds ratios (OR) with 95% CI were extracted from 36 studies. Statistical pooling of OR was performed using a random effect model with 95% CI. Forest plots were used to plot the effect size, either for each study or overall. We calculated the I2 statistics (0%~100%) to explain the between-study heterogeneity, with I2 ≤ 25% suggesting more homogeneity, 25% < I2 ≤ 75% suggesting moderate heterogeneity, and I2 > 75% suggesting high heterogeneity [15,16]. If the null hypothesis was rejected, a random effects model was used to calculate pooled effect estimations. If the null hypothesis was not rejected, a fixed effects model was used to calculate pooled effect estimations [14]; 95% CI was also reported. Forest plots were used to plot the effect size, either for each study or overall. Publication bias was evaluated by graphical inspection of funnel plot; estimation of publication bias was quantified by means of Egger's linear regression test [17]. In the case of moderate or high heterogeneity, influence analysis was performed with different approaches: Baujat plot [18] and a leave-one-out sensitivity analysis were performed by iteratively removing one study at a time to confirm that our findings were not driven by any single study. Metaregression analysis was performed, reporting results as regression coefficient (i.e., Beta) and p-value. One removed analysis was performed as a sensitivity analysis. "Meta package" in R-studio version 1. 1.463 (2009-2018) was used. Because this study was a systematic review and meta-analysis based on published articles, ethical approval was waived by the institutional review board of the University Hospital of Maastricht.

Study Inclusion
Our search yielded 877 records initially screened at the title and abstract level, with 222 papers fully reviewed for eligibility. There were no duplicate data. Ultimately, 36 studies were identified and provided data for the research analysis (Supplementary Figure S1). Table 1 shows the baseline characteristics of the included studies. The 36 studies included a total of 55,851 patients in the final analysis, from 2005 to 2018 . The mean age of the patients was 81.9 years, with a mean STS score of 8.3. Only five studies were prospective in nature [35,40,41,44,51]. The PPI details in the included studies are reported in Table 2. The overall incidence of PPI reached 15.2%, ranging from 4.3% to 32%. Values are n (%), mean = SD, or median (interquartile range) as appropriate. * Follow-up is reported as mean or median as given by the authors. LBBB = left bundle branch block; RBBB = right bundle branch block; SE = self-expandable; BE = balloon-expandable; ME = mechanically-expandable; TF = trans-femoral; TS = trans-subclavian; TA = trans-apical; T iliac = trans-iliac; T aortic = trans-aortic.

Discussion
Our meta-analysis demonstrates the impact of bundle branch block on post-TAVI PPI. Our study is derived from 36 studies reporting clinical outcomes in 55,851 patients receiving TAVI and presenting bundle branch conduction disturbances. The main results of this study can be assumed as follows: (i) the presence of LBBB does not influence post-TAVI PPI; (ii) the presence of RBBB has a significant impact on post-TAVI PPI.
The prevalence of LBBB in TAVI candidates is closed to 10% according to previous studies [45,55,56] and new-onset LBBB occurs in ∼ =30% after TAVI, depending on the type of prosthesis [9,57]. In the present study, the cumulative proportion of LBBB was 11.9%, including pre-operative LBBB and new-onset LBBB after TAVI. This rate is thus in accordance with previously published data [4,58]. The impact of LBBB on post-TAVI PPI remains controversial [49,50,59]. Data from the PARTNER experience [50,53] suggested that new-onset LBBB increases the rate of post-TAVI PPI. Moreover, an analysis at a national registry level as the prospective open TAVI Karlsruhe registry [48] identified slightly more PPI in patients with persistent new-onset LBBB. However, current data do not promote prophylactic PPI in patients presenting new-onset LBBB after TAVI [60] and some studies failed to identify LBBB as predictors for post-TAVI PPI [58,61]. In our study, the influence of LBBB on post-TAVI PPI was not significant. These controversial results can be related to various definitions of LBBB according to the degree of QRS prolongation defining the LBBB [58,[61][62][63] or to the high heterogeneity among the current studies [64]. Moreover, only three studies [22,50,53] quantified LBBB as left anterior fascicular or left posterior fascicular block, which may underestimate the impact of LBBB on post-TAVI PPI. As development of high-degree atrio-ventricular block is shown to be a common complication in patients with pre-existing LBBB (or RBBB) after TAVI [65], we shared the view of Waksman and colleagues [66] emphasizing the need for routine electrophysiology evaluation in patients undergoing TAVI in order to limit post-TAVI PPI. Electrophysiology testing differentiates the supra-nodal conduction disturbances from the infra-nodal ones by analyzing the atrial-His and the His-ventricular intervals. By integrating these values into the baseline electrocardiogram, we may be able to predict the evolution of the atrioventricular conduction in high-grade atrio-ventricular block. Such data may help the clinicians to choose between an aggressive strategy of early post-TAVI PPI or, on the contrary, an expectative strategy, avoiding some unnecessary post-TAVI PPI. The results of current ongoing trials investigating the use of electrophysiology-based algorithmic approaches with HV measurements as a guide for PPI, particularly in patients with LBBB (Clinical Monitoring Strategy vs. EP-Guided Algorithmic in LBBB Patients Post-TAVR (NCT03303612); The MARE Study (NCT02153307); Prospective Validation of a Pre-Specified algorithm for the management of conduction disturbances following transcatheter aortic valve replacement PROMOTE (NCT04139616)), are expected with impatience.
Pre-existing RBBB has been demonstrated in 6.9% [67] up to 13.6% [44] of patients undergoing TAVI and is a common underlying conduction disturbance in TAVI patients [10]. In this study, the cumulative proportion of RBBB among the included studies reached 9.2%, also in accordance with the studies of Auffret and colleagues [45], reporting a prevalence of RBBB of 10.3% in TAVI candidates. Several studies have already highlighted the role of pre-existing RBBB in the development of atrioventricular conduction, leading to post-TAVI PPI [2,3,22,26,33,39,41]. Our findings confirm the impact of such a condition on post-TAVI PPI, without the influence of age or LVEF. As previous studies emphasized the higher all-cause mortality and the poorer clinical outcomes in patients with pre-existing RBBB [10,45], bradyarrhythmia events may participate in the poorer outcomes of these patients. However, this analysis does not provide the cause of such discrepancies between RBBB and LBBB. The fragility of the right bundle branch during minor trauma or procedures may partially explain its highest incidence in the general population and, consequently, the more important impact of RBBB in post-TAVI PPI [68]. We look forward to having definitive data and guidelines to determine the optimal management and monitoring of patients with RBBB undergoing TAVI [60].
The current guidelines with respect to post-TAVI PPI [69] do not specify the adequate timing for post-TAVI PPI. Previously, the 2013 European Society Guidelines [70] recommended a clinical rhythm observation period of 7 days in the presence of high-grade trio-ventricular block before proceeding to post-TAVI PPI. In the current meta-analysis, the range of timing for PPI varied from day 0 to 1 year, also emphasizing the need for PPI after discharge and the influence of atrio-ventricular conduction in the long term [49]. This hypothetical observation period may be shortened or lengthened according to patient-related factors and electrical findings. Once again, electrophysiological study may identify some high-risk patients for the development of high-grade atrio-ventricular disturbances and, thereby, adjust the observation rhythm period in such patients [71].
As the definitive impact of infra-Hisian conduction disturbances on post-TAVI PPI has not definitively been clarified yet, specific recommendations for the type of valves used in such patients must be done with caution. However, some studies investigated the impact of the valves used on post-TAVI conduction disturbances [3,44,49,61]. Indeed, Franzoni and colleagues [61] found a higher incidence of LBBB when using a Medtronic CoreRevalving System with respect to the Edwards Sapien Valve. A more intra-ventricular implantation of the stent valve frame may result in higher rate of post-TAVI PPI [49]. Furthermore, the gap between the lower edge of the coronary cusp and the end of the frame of the prosthetics valve has been shown to be greater in patients with new-onset LBBB than in patients with post-operative conduction disturbances [61], emphasizing the role of the prosthesis placement in the development of post-operative conduction disturbances. Nevertheless, the requirement for post-TAVI PPI is multifactorial and studies incorporating electrophysiological findings with procedural data are mandatory [72].

Study Limitations
This meta-analysis has several limitations to be acknowledged. First, this metaanalysis possesses all the inherent bias associated with this investigation technique. Second, including pre-operative LBBB and new-onset LBBB in the analysis may be considered as a limitation for the interpretation of the final results, as the pre-operative LBBB can be related to other cardiovascular conditions and the new-onset LBBB can be related to procedural/interventional aspects. However, as our primary outcome was 30-day or inhospital PPI, we can afford to compare both LBBB as conduction disturbances present after TAVI, which can lead to PPI, independently of the chronic or acute apparition of the LBBB. Third, distinction of LBBB in left anterior fascicular or left posterior fascicular block was lacking in the majority of the included studies, thereby not providing enough information to perform in-depth analysis. Further research in this respect is currently needed. Finally, analysis of individual patient-level data may provide further understandings.

Conclusions
This meta-analysis found that patients with the presence of RBBB have higher risk for post-TAVI PPI, with no influence of age or LVEF. This finding has not been confirmed for patients experimenting with LBBB. Pre-operative evaluation strategies, including electrophysiological characteristics, are crucial in further extending patient selection for TAVI.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10 .3390/jcm10122719/s1, Figure S1. Flowsheet of the included studies, Figure S2. Forest plot pooling the proportion of LBBB in 33 studies, Figure S3. Forest plot pooling the proportion of LBBB in subset of 7315 patients with post-TAVI PPI. Figure S4. Forest plot pooling the proportion of LBBB in subset of 43,650 patients without post-TAVI PPI, Figure S5. Baujat plot: impact of the studies on overall heterogeneity in studies reporting on LBBB status, Figure S6. Bubble plots: influence of age on risk for post-TAVI PPI in patients with LBBB, Figure S7. Bubble plots: influence of LVEF on risk for post-TAVI PPI in patients with LBBB, Figure S8. Bubble plots: influence of the year of the study on risk for post-TAVI PPI in patients with LBBB, Figure S9. Forest plot pooling the proportion of RBBB in 28 studies, Figure S10. Forest plot pooling the proportion of RBBB in subset of 6932 patients with post-TAVI PPI, Figure S11. Forest plot pooling the proportion of RBBB in subset of 39,670 patients without post-TAVI PPI, Figure S12. Baujat plot: impact of the studies on overall heterogeneity in studies reporting on RBBB status, Figure S13. Bubble plots: influence of age on risk for post-TAVI PPI in patients with RBBB, Figure S14. Bubble plots: influence of LVEF on risk for post-TAVI PPI in patients with RBBB, Figure S15. Bubble plots: influence of the year of the study on risk for post-TAVI PPI in patients with RBBB. Institutional Review Board Statement: Because this study was a systematic review and metaanalysis based on published articles, ethical approval was waived by the institutional review board of the University Hospital of Maastricht.
Informed Consent Statement: Patient consent was waived due to the systematic review and metaanalytic nature of this study.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Conflicts of Interest:
The authors declare no conflict of interest.