Predictors of Permanent Pacemaker Implantation in Patients Undergoing Transcatheter Aortic Valve Replacement ‐ A Systematic Review and Meta‐Analysis

Abstract

Background

As transcatheter aortic valve replacement (TAVR) technology expands to healthy and lower‐risk populations, the burden and predictors of procedure‐related complications including the need for permanent pacemaker (PPM) implantation needs to be identified.

Methods and Results

Digital databases were systematically searched to identify studies reporting the incidence of PPM implantation after TAVR. A random‐ and fixed‐effects model was used to calculate unadjusted odds ratios (OR) for all predictors. A total of 78 studies, recruiting 31 261 patients were included in the final analysis. Overall, 6212 patients required a PPM, with a mean of 18.9% PPM per study and net rate ranging from 0.16% to 51%. The pooled estimates on a random‐effects model indicated significantly higher odds of post‐TAVR PPM implantation for men (OR, 1.16; 95% CI, 1.04–1.28); for patients with baseline mobitz type‐1 second‐degree atrioventricular block (OR, 3.13; 95% CI, 1.64–5.93), left anterior hemiblock (OR, 1.43; 95% CI, 1.09–1.86), bifascicular block (OR, 2.59; 95% CI, 1.52–4.42), right bundle‐branch block (OR, 2.48; 95% CI, 2.17–2.83), and for periprocedural atriorventricular block (OR, 4.17; 95% CI, 2.69–6.46). The mechanically expandable valves had 1.44 (95% CI, 1.18–1.76), while self‐expandable valves had 1.93 (95% CI, 1.42–2.63) fold higher odds of PPM requirement compared with self‐expandable and balloon‐expandable valves, respectively.

Conclusions

Male sex, baseline atrioventricular conduction delays, intraprocedural atrioventricular block, and use of mechanically expandable and self‐expanding prosthesis served as positive predictors of PPM implantation in patients undergoing TAVR.

Nonstandard Abbreviations and Acronyms

AR

aortic regurgitation

CVA

cerebrovascular accident

FU

follow up

HB

heart block

MACCE

major adverse cardiac or cerebrovascular event

MCRS

Medtronic Corevalve Revealing System

MR

mitral regurgitation

NOP LBBB

new onset persistent left bundle‐branch block

OCS

Observational Cohort Study

PPM

permanent pacemaker

RCT

randomized controlled trial

SAVR

surgical atrial valve replacement

SEV

self‐expandable valve

TAVR

transcatheter aortic valve replacement

VIV

valve‐in‐valve

Clinical Perspective

What Is New?

• This meta‐analysis comprising 78 studies (31261 patients) provides a comprehensive analysis of the predictors of pacemaker implantation in patients undergoing transcathether aortic valve replacement.

• Male sex, baseline atrioventricular conduction delays, and the use of mechanically expandable and self‐expanding prosthesis are associated with a higher need for permanent pacemakers after transcatheter aortic valve replacement.

What Are the Clinical Implications?

• Timely identification of these high‐risk patients can alleviate the risk of periprocedural atrioventricular block and associated complications such as syncope and sudden cardiac death.

As the rheumatic etiology of aortic stenosis (AS) has significantly waned over time, age‐related AS remains the most common valvular disease in the developed world.¹ Valve replacement is the only definite and effective treatment to improve survival in these patients, however, a multitude of coexisting comorbidities, including but not limited to chronic cardiac or pulmonary diseases, operative risks, extremes of age and poor physical health serve as barriers to surgical aortic valve replacement (SAVR). Transcatheter aortic valve replacement (TAVR) has recently emerged as a reasonable alternative to rescue these high‐risk patients.² The first TAVR was performed in 2002, in France, on a 57‐year‐old man in whom SAVR was contraindicated due to multiple comorbidities.³ Almost 20 years later, the data indicates that not only is it superior to medical therapy in patients with severe AS, but is also non‐inferior to SAVR, even in low‐risk patients.^{4, 5, 6}

However, like any other therapeutic intervention, the advent of TAVR has presented its own set of challenges urging the need for a favorable risk‐benefit estimation. With the widespread availability and expanded indication of TAVR to a lower‐risk healthy population, there are concerns about the rising trend of procedural complications associated with TAVR. A frequent issue encountered with this procedure is conduction defects requiring permanent pacemaker (PPM) implantation.^{7, 8} The bundle of His and the bundle branches run in the vicinity of where the prosthesis is being placed. These conduction abnormalities arise primarily due to the proximity of the aortic annulus to the atrioventricular conduction system that gets manipulated during the procedure.⁷ Data suggests that the prevalence of conduction defects post‐procedure also depends upon the type of valve implanted during the TAVR procedure.⁸ The 2 most common prostheses used are balloon‐expandable Edwards Sapien Valve (ESV) and self‐expanding Medtronic Corevalve Revealing System (MCRS) with a 5%–12% incidence of PPM implantation post‐procedure in the former and 24%–33% in the latter.⁹ Due to the manipulation of the old valve, aortic annulus dilatation and subsequent implantation of a prosthetic valve, conduction defects are common. In our study, we intend to identify various cardiac and non‐cardiac predictors that lead to PPM implantation following TAVR. We also aim to gauge the risk of conduction abnormalities based on the type of prosthesis and access site used in TAVR.

Methods

Data was obtained from published articles on the topic. All data can be obtained from the references mentioned in the supplementary file. The consolidated extracted data is available on demand.

Search Strategy

PubMed, Embase, Ovid, and Cochrane databases were queried with various combinations of keywords and medical subject headings (MeSH) to identify studies of interest. There were no time filters or language restrictions placed. Backward snowballing by screening the references of relevant articles were also performed to retrieve unidentified articles that were missed on the primary search. The MeSH used included 2 subsets: one for TAVR using the keywords “percutaneous prosthetic valve,” “transcatheter aortic valve replacement,” “TAVR,” “transcatheter aortic valve implantation,” “TAVI,” “percutaneous approach,” “minimal invasive aortic valve replacement,” “transapical aortic valve replacement,” and the other for PPM and heart block including “LAFB,” “LPFB,” “LBBB,” “pacemaker implantation,” “heart block,” “conduction abnormalities,” and “conduction delays.” The 2 subsets of MeSH were systematically combined using Boolean operators. The final results from all possible combinations were downloaded into an EndNote library. All randomized control trials (RCT) and observational cohort studies (OCS) until April 2021, were screened for relevance. Any OCS or RCT that assessed the post‐TAVR rate of atrioventricular conduction or cardiac rhythm abnormalities and subsequent PPM implantation during the same hospitalization or within 30‐days of TAVR procedure were included. To avoid the inclusion of duplicate data, we only selected the most contemporary data when overlapping study populations (according to the period of recruitment and participating institutions) were reported; however, we cautiously included all patients reporting different predictors from studies of overlapping populations. To measure the impact of the procedure on PPM implantation, all patients with prophylactic implantation of PPM before the TAVR procedure were excluded from the analysis.

Data Extraction

Raw data about the events of PPM implantation in different predictor comparison groups were extracted for analysis by the first 9 authors independently. Detailed study‐ and patient‐level baseline characteristics including the type of study design; recruitment period, region, and follow‐up duration; sample size, number of post‐TAVR PPM implantations, sex, age, procedural risk assessment (by logistic EuroSCORE [European System for Cardiac Operative Risk Evaluation] or STS‐PROM [Society of Thoracic Surgeons Predicted Risk of Mortality] score), and baseline comorbidities were abstracted. Additionally, data related to the access site (transfemoral versus trans subclavian, transapical versus transvascular), type of prosthesis (MCRS versus ESV versus LOTUS), inclusion criteria, and definition of outcomes were obtained from individual studies (Table S1). Finally, the post‐TAVR indications for PPM implantation in each article were also extracted. Based on previous reviews, the following proposed potential predictors were selected: age, sex, baseline conduction abnormalities, anatomical features, access route, and valve types. Case reports, review articles, conference papers, and articles with insufficient data or no control arms were excluded. Patients with prior PPM implantation unrelated to TAVR were also excluded from our analysis. All data was validated by the corresponding author; in case of missing data authors of the original article were contacted. The detailed search map is given in Data S1.

Statistical Analysis

The statistical analysis was performed using the DerSimonian and Laird (DL) and Mantel Haenszel (MH) methods on random‐ and fixed‐effects models, respectively. The unadjusted odds ratio (OR) for dichotomous outcomes of RCTs and OCS were calculated. The “test for overall effect” was reported as a z value corroborating the inference from the 95% confidence interval. To avoid the influence of study design on pooled estimates, a stratified analysis based on the type of study (OCS versus RCT) was performed. A subgroup analysis based on the type of implanted valve (mechanically expandable versus self‐expanding versus balloon‐expandable), access route (transfemoral versus trans subclavian), and procedure type (transapical versus transvascular) was also performed. Sensitivity analysis after exclusion of small studies with fewer than 200 patients was done to determine the impact of sample size on pooled estimates. Descriptive characteristics for continuous data were reported as mean and SD, whereas categorical variables were presented as frequencies and percentages. Higgins I‐squared (I²) statistical model was used to determine heterogeneity in outcomes of the included studies. The observed heterogeneity was regarded statistically significant if the I² statistics P value was <0.05. Publication bias was illustrated graphically using a funnel plot. The methodological quality assessment of the included RCTs was performed using the risk of bias‐2 (RoB‐2) tool and the Oxford quality scoring system (Jadad score). The Newcastle‐Ottawa Scale was used for assessing non‐randomized studies. The probability value of two‐sided P<0.05 was considered statistically significant. All statistical analysis was performed using the Cochrane Review Manager (RevMan) version 5.3 and STATA software (version 16.0, STATA Corp., College Station, Texas).

Quality of the Included Studies

The overall quality of the included studies was high. The risk of bias‐2 (RoB‐2) tool used 5 different bias assessments: selection, detection, performance, attrition, and reporting. All 3 of the included RCTs in our meta‐analysis were open‐label, posing some theoretical risk to “allocation concealment,” however, the overall risk of selection bias was reduced due to adequate randomization. Because most RCTs used an “intention to treat model” or had a lower loss at follow‐up, the risk of attrition bias was minimal. Similarly, the risk of reporting, detection and performance bias was lower due to appropriate reporting and adequate blinding of outcome assessors, respectively. The RoB‐2 plots are given in Figure 1. ^{10, 11, 12} The methodological quality of included RCTs was also high on the Jadad scale with a score >3 (Table S2). Observational studies were mostly matched in terms of clinical profile and demographics to curtail selection bias. The Newcastle‐Ottawa Scale for assessing nonrandomized studies indicated the inclusion of high‐quality observational studies (score >7) (Table S3).

Figure 1. Overall (A) and study‐level (B) methodological bias assessment of the included randomized clinical trials with the Cochrane risk of bias tool‐2.

 

Results

Search Results

The initial search revealed 4118 articles. After the removal of irrelevant (1561) and duplicate (2109) items, 448 studies were selected for full‐text review. Of these, 370 articles were excluded based on different reasons including: review articles (35), meta‐analyses (41), insufficient data for analysis (162), duplicate population studies (47), no risk factors data (80), and other reasons (5). A total of 78 articles (3 RCTs, 75 observational studies) qualified for quantitative analysis. The Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) flow diagram is shown in Figure 2 and the PRISMA checklist is given in Data S2.

Figure 2. PRISMA flow diagram showing the included studies.

 

Study Characteristics

A total of 31 261 patients undergoing TAVR from 78 studies were included, of these 6212 (19.8%) received PPM, while 25 049 (80.2%) did not require a PPM.^{7, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85} Most of the studies were from the United States and Europe. Two of the RCTs were multi‐continental, recruiting patients from the US, Australia, Germany, and Brazil. All included studies were published between 2009 and 2020 with an average recruitment period of approximately 4 years. The mean age of the included population was 81±8 years, comprising on average 46% male patients. The proportion of PPM implantation across different baseline comorbidities was comparable between the 2 groups. The detailed baseline characteristics are given in Tables S4 and S5, while the procedure characteristics of TAVR are given in Table S6. The summary is illustrated in Figure 1. The overall study‐level rate of post‐TAVR PPM ranged from 0.16% to 51.1%. The need for PPM implantation across different baseline comorbidities was variable as shown in Table S7 and Figure 3. The etiology for PPM implantation was only mentioned in 19.9% of patients (n=1238/6212). Post‐TAVR complete atrioventricular block was the most commonly observed indication for PPM implantation; other causes included bradycardia, new‐onset left bundle‐branch block (LBBB), and trifascicular block (Table). Patients with a prior history of PPM before the index TAVR procedure were mostly excluded from the analysis of their respective study. Two studies (De‐Carlo and Hamandi et al) had prophylactic PPM implantation before the TAVR procedure in 158 patients; these patients were excluded from the analysis. Most PPM implantations were performed during the same hospitalization or within 30‐days of the TAVR procedure. Most studies employed a transfemoral approach for TAVR, while 31 studies used transapical access in about 32% of its population. Mechanical (LOTUS) self‐expanding (MCRS and Evolut R) and balloon‐expandable (ESV) aortic prosthesis were the major valves used in the included studies. MCRS was used in 55, while ESV and Lotus were used in 46 and 12 studies, respectively. The mean log EuroSCORE for patients among the included studies was around 18.9±10 and the mean Society of Thoracic Surgeons score was found to be 5.85. The overall follow‐up duration ranged from 2 to 36 months, with a mean follow‐up of 8.02 months (Tables S4 through S6).

Figure 3. Percentages of patients with (A) different comorbidities and (B) those with and without permanent pacemaker (PPM) implantation across different baseline comorbidities.

A, Mean percentage of comorbidities. B, Proportion of comorbidities in PPM vs no‐PPM groups. ACS indicates acute coronary syndrome; AF; atrial fibrillation; OB, obesity; CABG, coronary artery bypass graft; CAD, coronary artery disease; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accidents; DM, diabetes; HLD, hyperlipidemia; HTN, hypertension; LD, Liver Disease; MI, myocardial infarction; PAD, peripheral arterial disease; PCI, percutaneous coronary intervention; SM, smoking. Asterisk denotes “prior history of”.

Table 1.

Periprocedural Causes of PPM Implantation in Patients Undergoing TAVR for Severe Aortic Stenosis

Periprocedural Events Leading to PPM in TAVR	No. of Patients	Percentage in the Known Causes
Third degree heart block	941	76%
LBBB	106	8.5%
Bradycardia	60	4.84%
Second degree AV block	45	3.63%
Second degree atrioventricular block associated with LBBB	36	2.9%
First degree atrioventricular block	35	2.82%
Tachy‐Brady syndrome	34	2.58%
Symptomatic pause	5	0.40%
Sick sinus syndrome	9	0.72%
Alternating RBBB and LBBB	4	0.32%
Afib with slow response	4	0.32%
Afib with complete atrioventricular block	4	0.32%
Total	1238	100%

All percentages are calculated among the known causes (1238). The reason for PPM implantation was not reported in 4924 cases. Afib indicates atrial fibrillation; LBBB, left bundle‐branch block; PPM, permanent pacemaker; RBBB, right bundle‐branch block; TAVR, transcatheter aortic valve replacement.

Pooled Analysis of Overall Studies

Twenty‐nine different potential predictors for the PPM implantation were evaluated. The number of patients having post‐TAVR PPM implantation (n=6212) from all studies contributed to the pooled OR calculation for each predictor. On a random effects model of binary data, the aggregate odds for post‐TAVR PPM implantation irrespective of the type of valve was higher in the male population compared with the female patients (OR, 1.16; 95% CI, 1.04–1.28). The baseline electrocardiographic conduction abnormalities, mobitz type‐1 second‐degree heart block (OR, 3.13; 95% CI, 1.64–5.93), mobitz type‐2 second‐degree heart block (OR, 3.89; 95% CI, 2.54–5.95), left anterior fascicular hemiblock (LAFB; OR, 1.43; 95% CI, 1.09–1.86), bifascicular block (OR, 2.59; 95% CI, 1.52–4.42), right bundle‐branch block (RBBB; OR, 2.48; 95% CI, 2.17–2.83), and intraprocedural atrioventricular block (OR, 4.17; 95% CI, 2.69–6.46) were associated with significantly higher odds of PPM implantation. The baseline predictor variables that were not statistically significantly associated with PPM implantation were age (OR, 1.19; 95% CI, 0.95–1.49), first‐degree heart block (OR, 1.09; 95% CI, 0.05–2.37), atrial fibrillation (AF; OR, 1.05; 95% CI, 0.93–1.20), left posterior fascicular hemiblock (LPFB; OR, 3.34; 95% CI, 1.1–11.13), left bundle branch block (LBBB; OR, 1.06; 95% CI, 0.87–1.29), severe pulmonary hypertension (OR, 1.78; 95% CI, 0.82–3.89), moderate/severe mitral regurgitation (MR; OR, 3.3; 95% CI, 0.59–18.32), unspecified heart failure; OR, 1.06; 95% CI, 0.72–1.55), and heart failure with preserved ejection fraction (OR, 1.01; 95% CI, 0.51–2.01). Of note, patients receiving 29 mm of prosthesis had significantly higher odds of PPM implantation compared with 23 mm prosthesis (OR, 1.49; 95% CI, 1.06–2.08). However, there appeared to be a statistically nonsignificant difference in the odds of PPM implantation between 23 mm versus 26 mm prosthesis (OR, 1.12; 95% CI, 0.62–2.03) and for patients with intraventricular septum size >11 mm (OR, 1.71; 95% CI, 0.17–17.41) and >22 mm (OR, 1.65; 95% CI, 0.55–4.93). The detailed valvular and anatomical variant estimates for PPM need are given Table S8.

Analysis of all predictors on a fixed‐effects model mirrored the findings of the random‐effects model with 2 exceptions; first‐degree heart block (OR, 0.35; 95% CI, 0.30–0.40) was found to be associated with a significantly lower risk, while LBBB (OR, 1.29; 95% CI, 1.14–1.46) had significantly higher odds of need for PPM. The detailed forest plots for both random and fixed effects are given in Figures S1 through S16. The heterogeneity in the outcomes of these studies was I²=0%, except for the studies comparing the RBBB and male populations, which showed significant heterogeneity (I²=52% and I²=74%, both P=<0.05), respectively (Figure 4). There was no significant difference in the odds of mortality in patients receiving PPM compared with those who did not receive PPM at 30 days and 1 year in 12 studies that included survival data (Figure 5).

Figure 4. Forest plot showing pooled estimates of demographic and electrocardiographic factors as potential predictors of permanent pacemaker (PPM) implantation in patients undergoing TAVR using random effects model (REM) and fixed effects model (FEM).

The dotted black line indicates null line (odds ratio of 1), to the right of null line indicates increased odds of PPM implantation. For each predictor the number of studies is in the parenthesis, the blue line indicates confidence interval, and the diamond red box signifies the point estimate. Lnor indicates natural log of odds ratio; lnlci, natural log of lower confidence interval; lnuci, natural log of uppper confidence interval.

Figure 5. Forest plot showing pooled estimates of procedural factors as potential predictors of permanent pacemaker (PPM) implantation in patients undergoing transcatheter aortic valve replacement (TAVR) using random effects model (REM) and fixed effects model (FEM).

 

On pooled analysis of continuous data, membranous septal length (MSL) was inversely, while the depth of prosthesis was directly, associated with the risk of PPM implantation. The mean MSL was 5.6 mm for patients requiring PPM implantation compared with 6.8 mm for those who did not require PPM, while the mean depth for prosthesis implantation for the former group was 6.86 mm compared with 5.34 mm in patients who did not require PPM (Figures S17 and S18).

Subgroup and Sensitivity Analyses

Overall, a head‐to‐head comparison based on the type of prosthesis favored the balloon‐expandable valves irrespective of the prevalence of different predictors. On a random‐effects model, the mechanically expandable valve (OR, 1.44; 95% CI, 1.18–1.76) and self‐expanding valves (OR, 1.93; 95% CI, 1.42–2.63) had higher PPM requirements compared with the self‐expanding and balloon‐expandable valves, respectively. Based on a breakdown data of 16 studies, MCRS implantation was associated with significantly higher odds of PPM implantation compared with ESV (OR, 2.48; 95% CI, 1.91–3.22). By contrast, the LOTUS valve implantation was associated with higher odds (OR, 1.61; 95% CI, 1.23–2.1) of PPM implantation compared with MCRS. Compared with EVOLUT‐R, the risk of PPM implantation was not significantly different in LOTUS and ESV (Table S9). There was no significant difference in the odds of PPM implantation in patients undergoing a transarterial versus transapical approach (OR 1.02; 95% CI, 0.1–10.1), transfemoral versus subclavian approach (OR 1.13; 95% CI, 0.6–2.1). These findings remained invariant on a fixed‐effects model. The heterogeneity among these studies ranged from I²=0% to I²=54% (Figure 6, Figures S14 through S16).

Figure 6. Forest plot showing the pooled estimate comparison of (A) self expanding vs balloon expandable and (B) mechanically expandable vs self‐expanding.

DL indicates DerSimonian and Laird; MH, Mantel‐Haenszel; PPM, permanent pacemaker.

Overall, a subgroup analysis based on the type of valve used, study design and access site mirrored the overall findings with few exceptions. In contrast to the pooled analysis, the summary estimates suggested higher odds of PPM implantation in patients with first‐degree heart block in MCRS (OR 1.95; 95% CI, 1.18–3.24). In concordance to the pooled analysis. Male sex (OR 1.33; 95% CI, 1.02–1.73), LAFB (OR 1.94, 95% CI, 1.11–3.38), intraprocedural atrioventricular block (OR 8.04; 95% CI, 3.53–18.29), and RBBB (OR 4.03; 95% CI, 2.47–6.56) remained the positive predictors of PPM implantation in a subset of patient undergoing MCRS‐only. For ESV and Evolut‐R valves, none of the previously mentioned predictors (except the intraprocedural atrioventricular block) appeared to have a significant influence on the need for PPM implantation. For individual valve types, we were able to assess only 5 to 10 predictors of PPM implantation (Table S9, Figures S19 through S21). More large scale studies are needed to determine the impact of other risk factors for PPM implantation across different valve types.

A sensitivity analysis on the “leave‐one‐out” strategy showed that the significantly lower odds of PPM implantation in patients with first‐degree heart block on a fixed‐effects model was driven by one study (Doshi et al) (Figure S22). There was no significant influence of any individual study on the pooled odds of PPM implantation across all other predictors (Figure S23 and S24). On a sensitivity analysis restricted to large studies of 200 patients or more, results remained consistent with the pooled results of the random‐effects model. Moreover, the summary estimates of OCS‐only (after exclusion of RCTs) and a subgroup analysis based on study design (OCS versus RCTs) also mirrored the results of the pooled analysis that included both OCS and RCT data (Table S10). The central illustration of all predictors is given in Figure 7 and the detailed study level PPM implantation rates for each predictor are given in Table S11.

Figure 7. Factors increasing the risk of PPM implantation post‐TAVR (red text indicates a higher risk).

 

Publication Bias

On the visual assessment of the funnel plots, no significant publication bias was detected for most of the predictors across all studies. Using the standard error, the vertical axis of the plot estimated the sample size of the study. Studies with a larger sample size were plotted on top and those with smaller populations appeared at the bottom of the plot. The horizontal spread indicated the individual effect size reflecting the overall power of the included studies. Our funnel plots were symmetrical, and most studies with low precision were spread evenly on both sides of the average line (Figure S25).

Discussion

The present meta‐analysis represents the most contemporary and largest evidence on the predictors of PPM implantation in patients with severe AS undergoing TAVR. Our findings revealed that male sex, pre‐TAVR baseline atrioventricular conduction abnormalities (including mobitz type‐1 second‐degree heart block, LAFB, RBBB), and intraprocedural atrioventricular block were associated with higher odds of PPM implantation, irrespective of the type of prosthesis or choice of the access site. A stratified analysis based on the prosthesis design showed a 2.4‐fold increased risk of PPM implantation with MCRS (self‐expanding) compared with ESV (balloon‐expandable), and 1.61 times higher odds of PPM‐need in LOTUS (mechanically expandable) compared with MCRS. The overall odds of PPM implantation remained identical in patients aged >80 years versus the younger population and those having first‐degree heart block, AF, prolonged PR‐interval, LPFB and LBBB, when compared with their corresponding control groups who had an absence of these rhythm abnormalities. The type of approach (transapical versus transvascular) or choice of access site (transfemoral versus trans‐subclavian) also had no impact on the risk of PPM implantation. Among the anatomical and valvular variants, the membranous septal length (MSL) was inversely, while the depth of prosthesis implantation was directly associated with the risk of PPM implantation. Larger devices (29 mm) had a higher risk of PPM implantation, while there was no impact of interventricular septum thickness, mitral regurgitation, or pulmonary hypertension on the need for PPM during TAVR. On subgroup analysis, only the MCRS data followed the results of the pooled analysis, indicating that the overall findings were mostly driven by the data obtained from patients receiving self‐expanding valves. The major post‐procedural etiology for PPM implantation was a periprocedural occurrence of high degree heart block, new‐onset LBBB, or persistent bradycardia.

It is imperative to identify patients at an increased risk of PPM implantation before a TAVR procedure, as timely detection of high‐risk patients can potentially prevent the occurrence of atrioventricular block and its associated complications (including syncope and sudden cardiac death). Also, patients with post‐TAVR atrioventricular nodal abnormalities are prone to prolonged hospitalization, putting a high financial burden on the healthcare budget.⁸⁶ PPM predictors in this context can help in the effective allocation of limited resources. With all its benefits, PPM placement comes at the cost of loss of atrioventricular synchrony, lack of physiological heart rate control, and increased risk of bleeding and pocket infection.^{87, 88} Early detection of patients at high risk of PPM implantation and identification of pre‐specified predictors, therefore provides an opportunity to mitigate these risks and to favorably lower the harm‐benefit ratio.

Among the measured predictors for PPM implantation, the demographic risk factors including age and sex are of paramount importance. Current evidence on sex‐related differences in post‐TAVR complications and the need for PPM is conflicting in recently published studies.^{89, 90}Our large‐scale analysis shows a 16% higher rate of PPM implantation in men. This can partly be explained by the relatively larger‐sized bioprosthesis (>25 mm) they receive, but mostly because of the higher prevalence of baseline comorbidities, putting men at a greater risk of procedural complications.^{63, 90} Additionally, our results also revealed a numerically higher rate of PPM use (by 19%) in a population age >80 years, however, the difference did not reach statistical significance. These findings contrast the results of Ramkumar et al. and Ledwoch et al. studies, which denoted a significantly higher risk of post‐TAVR PPM placement in octogenarians by 30% and 35%, respectively.^{37, 44} Amongst the cardiac predictors, the presence of a LAFB, bi‐fascicular block and second degree atrioventricular block are known to be associated with higher chances of receiving a PPM after TAVR.^{7, 8, 9} Our study echoes the same trend and expands these findings by demonstrating a 1.3‐, 2.1‐, and 3.1‐fold increase in the odds of the need for PPM implantation in LAFB, bi‐fascicular block and second degree atrioventricular block, respectively.⁹ Regarding the baseline first‐degree atrioventricular block, Dolci et al and Naveh et al showed an increased incidence of PPM placement at 1 year of TAVR.^{46, 91} By contrast, we believe that a first‐degree atrioventricular block is a mere delay of atrioventricular conduction rather than a true block and that is why our study demonstrated no impact of first‐degree heart block on the need for PPM implantation.

Studies have shown a higher incidence of post‐TAVR atrioventricular blocks in patients with baseline conduction blocks, due to the manipulation of an already diseased conduction system.^{37, 44, 46, 61, 73, 92, 93} Pre‐procedure LBBB and RBBB resulted in up to 1.5 times greater risk of PPM implantation after TAVR.^{92, 93} In our study, RBBB conferred a 2.48 times greater risk of PPM implantation, much higher than the expected rise seen in previous studies. Intriguingly, baseline LBBB on our analysis did not increase the peri‐procedural odds of atrioventricular block or the need for PPM implantation on a random‐effects model. These effects were consistent across the different types of prosthesis and access sites used for the TAVR procedure. When comparing the risk of atrial arrhythmias induced conduction abnormalities, we found that AF had no impact on the need for PPM implantation after TAVR. These findings were in line with the previous literature that also demonstrated an identical rate of need for PPM.⁹⁴ While a subset of the PARTNER registry showed that patients with sinus rhythm before TAVR and AF at discharge were twice more likely to get a PPM, patients with chronic AF had <6% risk of PPM, not significantly different from patients having no‐AF at baseline.⁹⁴

On review, we found 40 previous meta‐analyses discussing the risk factors of PPM implantation, however in light of the current evidence the applicability of those studies is limited.^{92, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132} Most of these meta‐analyses included a smaller number of previously published studies ranging from 4 to 41 articles, missing a large amount of contemporary data. The selection criteria and measured predictors were limited with respect to conduction abnormalities evaluated, indications for TAVR, and in some incidences inclusive of SAVR patients. More importantly, these studies had conflicting results. By contrast, our meta‐analysis is the largest study (78 studies), including all patients who underwent TAVR for symptomatic AS (irrespective of the etiology), a wider range of demographics predictors, conduction abnormalities and procedural characteristics (29 predictors). Our study also provides a subgroup analysis on the type of valve and sensitivity analysis based on the sample size and study design. The detailed study‐level characteristics and differences of our study from previous meta‐analyses are given in Table S12.

Previous small‐scale studies have also shown that atrioventricular conduction disturbances and a subsequent requirement for PPM were more common after the implantation of non‐balloon expandable valves.^{111, 133} Our results validated these findings by demonstrating a 1.93 and 2.8 times higher rate of PPM implantation in the self‐expanding and mechanically expandable prosthesis compared with the balloon‐expandable valves. MCRS and LOTUS, being a self‐expanding and mechanically expandable valve increases the risk of complete heart block due to deeper implantation into the aortic annulus, tissue edema, and sustained pressure on the conduction pathway (atrioventricular node and left bundle branches).⁷³ These effects might be delayed in the balloon‐expandable valves (ESV) due to the intermittent nature of expansion and lower risk of tissue impingement. Although relatively lower, the newer generation balloon‐expandable prosthesis is not devoid of the risk of PPM implantation. A study by Bisson and colleagues noted that in an effort to decrease a paravalvular leak, the newer ESV comes with an outer skirt, increasing the odds of PPM implantation.¹³⁴ In contrast to the studies by Puls et al and Rouge et al that showed a higher prevalence of PPM implantation in transfemoral approach compared with trans subclavian access, we found no impact of the choice of the TAVR access site (transapical versus transvascular) and (transfemoral versus trans subclavian) on the need for PPM implantation.^{38, 135} To summarize, men, patients with baseline conduction abnormalities and those receiving the self‐expanding or mechanically expandable prosthesis are at higher risk of PPM implantation after TAVR.

Limitations

Our study is constrained by the limitations of the included studies. A multivariate logistic regression model is required to control for potential confounders and to obtain an independent impact of the predictor. Patient‐level data were missing to determine the adjusted odds of PPM predictors. For the same reason, we could not assess the impact of the procedure technique and could not account for the differential use of medications or other causes of atrioventricular conduction abnormalities. The impact of unmeasured confounding factors and operators' skills could not be measured. Although we selected a wide range of potential, previously proven predictors, the available data for some comparisons were sparse. Due to the lack of extended follow‐up data the long‐term effectiveness of PPM could not be evaluated. It is also important to note that the reasons for PPM implantation were variable in included studies, hence PPM implantation in our analysis should not be interpreted as a surrogate marker of atrioventricular conduction disturbances. The need for PPM in post‐TAVR patients can be influenced by several economic and logistic factors out of the scope of the current study.

Conclusions

Patients with baseline conduction abnormalities, men, and those receiving mechanical‐ or self‐expanding larger‐sized prostheses for transcatheter aortic valve replacement are at an increased risk of pacemaker implantation. Given the clinical and economic impact of TAVR, interventionists should cautiously risk‐stratify and identify patients at a high risk of the need for PPM.

Sources of Funding

None.

Disclosures

None.

Supporting information

Data S1

Tables S1–S12

Figures S1–S25

References 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135