One-year pacing dependency after pacemaker implantation in patients undergoing transcatheter aortic valve implantation: Systematic review and meta-analysis.

Objectives: Atrioventricular conductions disturbances, requiring permanent pacemaker implantation (PPI), represent a potential complication after transcatheter aortic valve implantation (TAVI), However, little is known about the pacemaker dependency after PPI in this patient setting. This systematic review analyses the incidence of PPI, the short-term (1-year) pacing dependency, and predictors for such a state after TAVI.
Methods: We performed a systematic search in PUBMED, EMBASE, and MEDLINE to identify potentially relevant literature investigating PPI requirement and dependency after TAVI. Study data, patients, and procedural characteristics were extracted. Odds ratio (OR) with 95% confidence intervals were extracted.
Results: Data from 23 studies were obtained that included 18,610 patients. The crude incidence of PPI after TAVI was 17% (range, 8.8%-32%). PPI occurred at a median time of 3.2 days (range, 0-30 days). Pacing dependency at 1-year was 47.5% (range, 7%-89%). Self-expandable prosthesis (pooled OR was 2.14 [1.15-3.96]) and baseline right bundle branch block (pooled OR was 2.01 [1.06-3.83]) showed 2-fold greater risk to maintain PPI dependency at 1 year after TAVI. Conclusions: Although PPI represents a rather frequent event after TAVI, conduction disorders have a temporary nature in almost 50% of the cases with recovery and stabilization after discharge. Preoperative conduction abnormality and type of TAVI are associated with higher PPI dependency at short term.

Rate of pacemaker dependency across the time after TAVI.

Up to 50% of the patients with permanent pacemaker implantation following TAVI exhibit no pacemaker dependency at 1-year follow-up.

Better understanding of pacemaker dependency after TAVI should allow better oriented guidelines with respect to indications and timing of pacemaker implantation, as well as postpermanent pacemaker implantation management, based on the high recovery rate of effective native atrio/ventricular conduction.

See Commentaries on pages 56 and 58.

Transcatheter aortic valve implantation (TAVI) was first introduced in 2002 as a less-invasive therapeutic option in patients with severe symptomatic aortic stenosis unfit for cardiac surgery. Nowadays, there is a trend to extend these procedures even to intermediate- and low-risk patients, making the frequency of TAVI procedures grow exponentially. However, complications, such as atrioventricular conduction disturbances requiring permanent pacemaker implantation (PPI), may diminish the benefit of these procedures: the incidence of PPI after TAVI, for instance, represents a rather frequent event after TAVI. Indeed, an association between PPI and all-cause deaths and heart failure rehospitalizations at 1 year from TAVI has been recently shown. However, a certain percentage of atrioventricular conduction abnormalities after TAVI may resolve following PPI, even after a few days from implant. Presently, the current literature available on pacing dependency after TAVI is limited and based on studies with small patient samples (evidence level B)., Therefore, we performed this systematic review to determine the incidence of PPI, the pacing dependency, and potential predictors for pacing-dependency at 1 year after TAVI procedures.

Methods

Research Strategy

A broad, computerized literature search was performed to identify all relevant studies from PubMed, Embase, and MEDLINE databases. 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 implantation. We restricted the research to English-language publications. Last access to the database was on April 25, 2020. The search was limited to studies in human recipients. A framework of the systematic review process is plotted in Figure 1, in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 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.

Figure 1

Study selection. Flow diagram of included studies based on the Preferred Reported Items for Systematic Reviews and Meta-Analysis (PRISMA).

Eligibility Criteria and Study Selection

Studies were included in the final analyses if patients were (1) >18 years; (2) >250 patients were included in the main analysis, to provide data interpretation of the most consistent clinical series; and (3) studies provided a definition of cardiac pacemaker dependency. Other studies following the same criteria, but having a smaller patient sample size (<250 patients), were included in a separated analysis (secondary analysis) and are provided in Appendix E1. Multiple publications from a single center were managed to include the last publication. If the same authors published more studies with the same series, the largest patient cohort was included. Studies were excluded if 1 of the following criteria was present: (1) presence of congenital pathology; (2) patients undergoing noncardiac surgery procedures or transcatheter procedure or heart transplantation; (3) no information provided about PPI; (4), publication before year 2002; or (5) outcomes not clearly reported or impossible to extract or calculate from the available results. Review, clinical update, and case reports were not taken into account. All potentially relevant studies were reviewed in detail to check their adhesion to the inclusion criteria. Title and abstracts of all retrieved paper were independently reviewed by 2 researchers (J.R. and M.D.M.) to identify studies fulfilling the inclusion criteria. Controversial findings were solved by the intervention of a third reviewer (R.L.). The quality of included studies was assessed using the Newcastle–Ottawa Scale for observational studies by 2 investigators independently (J.R. and M.D.M.).

Data Extraction

Microsoft Office Excel 2016 (Microsoft, Redmond, Wash) was used for data extraction that was performed independently by 2 researchers (K.V., L.V.). Year of publication, study design, sample size, age, Society of Thoracic Surgeons (STS) score, inclusion period, left ventricular ejection fraction, peripheral vascular disease, diabetes mellitus, valve type, follow-up, approach for TAVI, indications for PPI, timing of PPI (days), PPI rate, dependency definition, dependency follow-up (months), multivariable predictors of PPI, and PPI-related complications were extracted.

Statistical Analysis

The primary end point was 1-year pacing dependency, defined in different ways. Calculation of proportions of PM dependency at different time points was obtained using a meta-analytic approach by means of metaprop function of meta package in R (Foundation for Statistical Computing, Vienna, Austria). Odds ratio (OR) with 95% confidence intervals (CIs) were extracted. 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. If the null hypothesis was rejected, a random effects model was used to calculate pooled effect estimates. If the null hypothesis was not rejected, a fixed-effects model was used to calculate pooled effect estimates; 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 performed with trim-and-fill method and quantified by means of Egger's linear regression test. On-leave out study analysis was performed as sensitivity analysis in case of moderate or high heterogeneity. Meta-regression was performed to test the influence of age, time of pacemaker implantation, and sample size. The software used for the analyses was R-studio (meta package), version 1.1.463 (2009-2018).

Results

Study Selection

A total of 877 records were initially screened at the title and abstract level, with 801 papers fully reviewed for eligibility. There were no duplicate data. Ultimately, 23 studies were identified and provided data for the research analysis.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 An additional 30 studies are considered in Appendix E1 because of their limited sample size (<250 patients). The flowsheet for selection of included studies is represented in Figure 1. The Newcastle-Ottawa Scale confirmed a high-quality level for all studies included in the main analysis (Table E1).

Table E1

Newcastle–Ottawa Scale for the assessment of the risk of bias in individual nonrandomized studies

Author	Score	Selection	Comparability	Outcome/exposure
Bjerre Thygesen et al9	8	∗∗∗∗	∗∗	∗∗
Urena et al10	9	∗∗∗∗	∗∗	∗∗∗
Nazif et al11	9	∗∗∗∗	∗∗	∗∗∗
Van Gils et al12	9	∗∗∗∗	∗∗	∗∗∗
Raelson et al13	8	∗∗∗∗	∗∗	∗∗
Dumonteil et al14	9	∗∗∗∗	∗∗	∗∗∗
Kaplan et al15	9	∗∗∗∗	∗∗	∗∗∗
Chamandi et al16	8	∗∗∗∗	∗∗	∗∗
Gaede et al17	8	∗∗∗∗	∗∗	∗∗
Gonska et al18	7	∗∗∗∗	∗	∗∗
Marzahn et al19	9	∗∗∗∗	∗∗	∗∗∗
Nadeem et al20	9	∗∗∗∗	∗∗	∗∗∗
Campelo-Parada et al21	8	∗∗∗∗	∗∗	∗∗
Mirolo et al22	8	∗∗∗∗	∗∗	∗∗
Van Gils et al23	8	∗∗∗∗	∗∗	∗∗
Takahashi et al24	8	∗∗∗∗	∗∗	∗∗
Chan et al25	9	∗∗∗∗	∗∗	∗∗∗
Ghannam et al26	9	∗∗∗∗	∗∗	∗∗∗
Costa et al27	9	∗∗∗∗	∗∗	∗∗∗
Dolci et al28	9	∗∗∗∗	∗∗	∗∗∗
Tovia-Brodie et al29	9	∗∗∗∗	∗∗	∗∗
Junquera et al30	9	∗∗∗∗	∗∗	∗∗∗
Meduri et al31	9	∗∗∗∗	∗∗	∗∗∗

Study, Participants, and Procedural Characteristics

Baseline characteristics of selected studies are represented in Table 1. Studies were published between 2014 and 2019, and patient recruitment occurred between January 2005 and February 2018; 26% of them (n = 6) were prospective,,,,,, and the remaining studies were all retrospective observational. Mean duration of follow-up was 16 months.

Table 1

Baseline characteristics of studies including >250 patients (n = 23)

Study	Year	Study design (no. centers)	Sample size	Age, y	STS score, %	Inclusion period	Left ventricle ejection fraction, %	Peripheral vascular disease, %	Diabetes mellitus, %	Valve type	Follow-up, mo∗	Approach for TAVI	Mortality at 30 d
Bjerre Thygesen et al9	2014	Prospective (1)	258	na	na	na	na	na	na	100% SE	na	na	na
Urena et al10	2014	Retrospective (8)	1556	80.5	7.5	January 2005 to February 2013	55.5	na	31.1	55.1% BE44.9% SE	22	na	7%
Nazif et al11	2015	Retrospective (21)	1973	84.5	11.4	May 2007 to September 2011	53.7	42.4	36	100% BE	12	na	6.6%
Van Gils et al12	2017	Retrospective (4)	306	83	6.3	May 2008 to February 2016	na	22	30	38.2% SE34.7% SE27.1% ME	12	na	7%
Raelson et al13	2017	Retrospective (1)	578	85.5	na	March 2009 to December 2014	na	na	na	21% SE79% BE	1	na	na
Dumonteil et al14	2017	Retrospective (14)	250	84	6.3	October 2012 to May 2014	58.3	na	22.5	100% ME	12	100% Transfemoral	4%
Kaplan et al15	2019	Retrospective (1)	594	81.6	na	January 2011 to December 2017	na	na	na	na	12	na	na
Chamandi et al16	2018	Prospective (9)	1692	81.5	10.9	May 2009 to February 2015	na	na	33.1	50.3% BE49.7% SE	48	na	42.3%
Gaede et al17	2018	Retrospective (1)	1025	81.9	na	2010-2015	na	21.1	33.3	na	2.4	na	na
Gonska et al18	2018	Prospective (1)	612	80.4	6.5	February 2014 to September 2016	57.5	na	29.9	58.8% BE36.7% ME4.4% SE	12	na	1.3%
Marzahn et al19	2018	Retrospective (1)	856	80.5	na	July 2008 to May 2015	57.5	na	38	37.4% SE57.8% BE4.8% ME	12	100% Transfemoral	na
Nadeem et al20	2018	Retrospective (1)	672	81.4	7.4	2011-2017	53.8	21.5	41	na	12	na	na
Campelo-Parada et al21	2018	Retrospective (2)	347	na	na	May 2010 to December 2015	54.6	na	37.3	100% BE	1	77.8% Transfemoral	na
Mirolo et al22	2018	Retrospective (1)	936	na	na	October 2009 to January 2017	na	na	na	95% BE5% SE	2.3	na	na
Van Gils et al23	2018	Retrospective (1)	291	79	na	January 2012 to December 2015	na	na	34	42% SE51.5% BE6.5% ME	12	94% transfemoral6% transsubclavian	5%
Takahashi et al24	2018	Retrospective (4)	1621	84.3	na	January 2010 to December 2014	na	na	na	72.5% SE	13	na	na
Chan et al25	2018	Retrospective (2)	913	81.6	na	January 2012 to December 2017	na	na	na	na	12	na	na
Ghannam et al26	2019	Retrospective (1)	573	79.8	6.4	January 2012 to September 2017	57.2	45.8	37.8	100% SE	17.1	na	na
Costa et al27	2019	Prospective (1)	1116	82	4.4	June 2007 to February 2018	53.3	na	28.9	61.8% SE27.2% BE0.5% ME10.5% Others	72	97% transfemoral3% others	3.9%
Dolci et al28	2019	Retrospective (1)	266	80	na	February 2014 to February 2018	53	22	28	100% BE	12	84% transfemoral16% transapical	na
Tovia-Brodie et al29	2019	Prospective (1)	795	82.5	na	April 2012 to December 2016	na	na	na	na	28.2	na	na
Junquera et al30	2019	Retrospective (2)	676	82	5	May 2007 to March 2017	57.4	na	31.2	60.5% BE35.2% SE0.7% ME0.3% others	12	64.8% transfemoral	na
Meduri et al31	2019	Prospective (1)	704	82.5	6.6	na	na	28.2	30.9	34% SE66% ME	12	na	na
Total	na	79	18,610	81.8 (43-102)	7.1 (0.74-34)	January 2005 to February 2018	55.6	29	32.6	SE 38%BE 46%ME 13.5%others 2.5%	16	88.2% transfemoral2.3% transapical0.9% transsubclavianOthers 8.7%	9.6%

Values are n (%). STS score, Society of Thoracic Surgeons Risk Score; TAVI, transaortic valve implantation; na, not available; SE, self-expandable; BE, balloon-expandable; ME, mechanically expandable.

Follow-up is reported as mean or median as given by the authors.

In total, 18,610 patients were included in 23 studies from 79 centers. The crude incidence of PPI after TAVI procedure was 17%, ranging from 8.8% to 32%. Preprocedural risk was assessed by the STS score in the majority of the studies (11/23) and the mean STS score was 7.1 (0.74-34). Mean age was 81.8 years (range, 43-102 years). The implanted device was a self-expandable (SE) valve in 38% of patients, a mechanically expandable valve in 13.5%, and a balloon-expandable (BE) in 46% of the patient population. Others devices were used in 2.5% of the patients. Three studies used only BE valves,,, and 3 studies implanted solely SE valve.,, The transfemoral access route was preferred (88.2%) over the other approaches (transapical 2.3%; transsubclavian 0.9%, and other approach used in 8.7%). The mortality at 30 days was 9.6%. The baseline characteristics of the studies including <250 patients are provided in Table E2.

Table E2

Baseline characteristics of studies including <250 patients (n = 30)

Study	Year	Study design (no. centers)	Sample size	Age, y	STS score, %	Inclusion period	Left ventricle ejection fraction, %	Peripheral artery disease, %	Diabetes mellitus, %	Valve type	Follow-up, mo∗	Approach for TAVI	Mortality at 30 d
Sinhal et alE1	2008	Prospective (1)	106	84.2	na	na	62.5	na	na	100% SAPIEN	1	na	na
Jilaihawi et alE2	2009	Retrospective (1)	30	84.4	8.3	January 2007 to March 2008	47.5	na	17.6	100% MCV	na	na	8.8%
Baan et alE3	2010	Retrospective (1)	34	80.4	5	na	na	na	32	100% MCV	1	na	20.5%
Fraccaro et alE4	2011	Retrospective (1)	64	81	na	May 2007 to April 2009	52.3	34.4	na	100% MCV	12	94% transfemoral6% transsubclavian	5%
Van der boon et alE5	2013	Prospective (1)	167	81	na	November 2005 to February 2011	51	10.2	21.6	100% MCV	11,5	97% transfemoral3% transsubclavian	na
Pereira et alE6	2013	Retrospective (1)	65	79.3	na	August 2007 to May 2011	na	47.7	38.5	100% MCV	6	78.5% transfemoral20% transsubclavian1.5% direct aortic	na
Goldenberg et alE7	2013	Retrospective (1)	191	na	na	February 2009 to July 2012	na	na	na	65% MCV35% SAPIEN	17	na	na
Ramazzina et alE8	2014	Retrospective (1)	97	83	na	October 2010 to January 2013	55	na	26.4	61% MCV39% SAPIEN	12	100% transfemoral	na
Boerlage-Van Dijk et alE9	2014	Retrospective (1)	121	80.5	4.5	October 2007 to June 2011	na	na	28	100% MCV	12	100% Transfemoral	na
Renilla et alE10	2015	Retrospective (1)	95	na	na	January 2007 to December 2011	na	na	na	100% MCV	35	86.9% transfemoral13% transsubclavian	na
Petronio et alE11	2015	Prospective (9)	194	80.2	7.2	October 2011 to April 2013	na	27.6	31	100% MCV	1	89.7% transfemoral6.2% transsubclavian4.1% direct aortic	1.6%
Weber et alE12	2015	Retrospective (1)	212	80.8	9.4	2008 - 2012	52.8	na	25	100% MCV	9	100% transfemoral	6.1%
Schernthaner et alE13	2016	Retrospective (1)	153	81	6	na	na	na	31	82% MCV18% SAPIEN	1,5	94% transfemoral6% direct aortic	na
Kostopoulou et alE14	2016	Prospective (1)	45	81	na	January 2010 to February 2012	49	na	27	100% MCV	24	na	na
Sideris et alE15	2016	Prospective (1)	168	na	na	January 2009 to October 2015	na	na	na	100% MCV	na	100% transfemoral	na
Luke et alE16	2016	Retrospective (1)	140	81	na	July 2011 to May 2016	na	na	na	81% MCV19% EVOLUT	na	100% transfemoral	na
Makki et alE17	2017	Retrospective (1)	172	83	na	November 2011 to January 2016	51	na	46	92% MCV8% LOTUS	22	100% transfemoral	na
Nijenhuis et alE18	2017	Retrospective (1)	155	80.5	6	June 2007 to June 2015	59	30.3	34.8	na	24	na	na
Naveh et alE19	2017	Prospective (1)	110	80.7	na	September 2008 to November 2013	57.6	23.6	30	75.5% MCV24.5% SAPIEN	12	88.2% transfemoral6.4% transsubclavian5.5% direct aortic	na
Alasti et alE20	2018	Prospective (1)	152	83.6	na	April 2012 to October 2016	59.2	5.3	18.7	100% LOTUS	12	99.4% transfemoral0.6% direct aortic	2.6%
Ortak et alE21	2018	Prospective (1)	66	80.4	3.7	2014 -2016	53.2	na	na	100% LOTUS	7	na	3.5%
Rodes-Cabau et alE22	2018	Prospective(11)	103	80	5	June 2014 to July 2016	56	na	43	14.5% MCV37% EVOLUT R51.5% SAPIEN	12	86% transfemoral10% transapical4% transsubclavian	na
Leong et alE23	2018	Retrospective (1)	67	80.5	na	January 2013 to December 2015	na	na	30	16.4% EVOLUT R35.8% MCV34.3% SAPIEN13.4% Others	2,4	na	na
Sharma et alE24	2018	Prospective (1)	226	81.2	na	March 2012 to October 2016	na	na	na	100% BE	1	na	na
Bacik et alE25	2018	Prospective (1)	116	77.1	na	August 2013 to Mar 2017	50.5	na	40.5	100% SAPIEN	12	82.8% transfemoral17.2% transapical	na
Megaly et alE26	2019	Prospective (1)	172	na	na	January 2010 to May 2017	na	na	50	na	12	na	na
Yazdchi et alE27	2019	Retrospective (1)	na	na	na	2013 - 2017	na	na	na	na	14	na	na
McCaffrey et alE28	2019	Retrospective (1)	98	79.6	5	May 2015 to March 2018	55.7	26	35	100% SAPIEN	1	93% transfemoral6% transapical1% direct aortic	4.8%
Miura et alE29	2019	Retrospective (1)	201	84.8	6.4	October 2013 to September 2016	60.6	9	26	100% SAPIEN	13.5	68% transfemorall 27% transapical5% transiliac	0.5%
Dhakal et alE30	2020	Retrospective (1)	176	80	5.7	Seprember 2012 to March 2017	53	31	43	na	18.9	na	na
Total	na	49 centers	3612	81.2	6	November 2005 to March 2018	54.4	24.5	32.1	65.6% MCV23.1% SAPIEN8% LOTUS2.7% EVOLUT0.5% Others	11.4	92.5% transfemoral3% transsubclavian1% direct aortic3.2% transapical0.3% transiliac	6%

Values are n (%). STS score, Society of Thoracic Surgeons Risk Score; TAVI, transaortic valve implantation; na, not available; MCV, Medtronic CoreValve.

Follow-up is reported as mean or median as given by the authors.

PPI Details

The PPI details in studies including >250 patients are reported in Table 2. Data regarding PPI details in studies including <250 patients are described in Table E3. All studies provided indications for PPI, except in 4 series.,,, The main indications for PPI were divided into 3 main categories of atrioventricular conduction disorders: atrioventricular block (second and third degree) in 82.7% of the patients, sick sinus syndrome in 2.7%, severe bradycardia in 2.8%, and others indications in 11.8% of the patients, respectively. The timing of PPI varied remarkably among studies, occurring at a median time of 3.2 days (range, 0-30 days) from TAVI procedure. Nine studies didn't provide information about the timing of PPI.,18, 19, 20, 21,,,, The overall incidence of PPI reached 17% of the TAVI patients, ranging from 8.8% to 32% of the cases.

Table 2

Pacemaker-related details in studies including >250 patients (n = 23)

Study	Indications for PPI	Timing of PPI, d∗	PPI rate	Dependency definition	Dependency rate	Dependency follow-up, mo∗	Multivariable predictors of PPI	Association	PPI-related complications
Bjerre Thygesen et al9	100% AVB	na	27.4%	Resolution of conduction abnormalities	50%	na	na	na	na
Urena et al10	75.3% AVB7.1% SSS7.9% bradycardia9.6% others	3	15.4%	“paced rhythm” reported	66.9%	12	na	• PPI protective factor for the occurrence of unexpected (sudden or unknown) death • Negative effect on left ventricular function over time	na
Nazif et al11	79% AVB17.3% SSS	3	8.8%	“ventricular pacing” reported	50.5%	12	• Pre-existing RBBB • Prosthesis to left ventricle outflow tract diameter ratio • Left ventricle -end diastolic diameter	• Longer duration of hospitalization • Greater rates of repeat hospitalization and mortality or repeat hospitalization at 1 y	na
Van Gils et al12	99% AVB1% SSS	2	41%	% ventricular pacing rhythm reported	89%	12	• LOTUS valve • Greater BMI before TAVI	• RBBB at baseline associated with greater PPI	na
Raelson et al13	82% AVB	3	9%	No intrinsic ventricular activity during pacing at 30 bpm	39%	1	na	na	na
Dumonteil et al14	88.9% AVB5.9% others	3	32%	“paced rhythm” reported	55.4%	12	• Baseline RBBB • Left ventricle outflow tract overstretch >10%	• Trend lower PPI rate at 30 d with shallower (≤5 mm) implant depth	na
Kaplan et al15	79% AVB21% others	2,5	13.1%	High-grade AVB with a ventricular escape rate of less than 40 beats/min on device interrogation	21.9%	12	na	• Use of SE valves and postballoon dilatation associated with markedly increased risk of PPM dependency	na
Chamandi et al16	76.7% AVB5.6% SSS3.1% Bradycardia14.6% others	2	19.8%	100% right ventricular pacing	27.4%	48	na	• PPI greater rates of rehospitalization due to heart failure and combined end point of mortality or heart failure rehospitalization • PPI lesser improvement in left ventricle ejection fraction over time, particularly if reduced before TAVI	na
Gaede et al17	90% AVB8% SSS2% Bradycardia	4	14.7%	Ventricular pacing >95%	29.5%	2.4	• Pre-existing RBBB • CoreValve prosthesis	Predictors of lack of recovery of AVB• Previous RBBB • Greater mean aortic valve gradient • Postdilatation of the prosthesis	na
Gonska et al18	85.% AVB10.1% Bradycardia4.8% others	na	24.4%	“ventricular pacing” reported	30.9%	1	na	• PPI without significant impact on survival or combined end point of major adverse events within 1 y	1.8% reoperation due to lead dislocation2.4% hematomas/bleeding at the site of the pacemaker
Marzahn et al19	89% AVB5.5% Bradycardia4.1% SSS1.4% others	na	16.9%	“right ventricular pacing %” reported	55%	12	na	na	na
Nadeem et al20	na	na	21.7%	“right ventricular pacing %” reported	45.5%	12	na	• PPI more likely to have heart failure admissions • PPI trend toward increased mortality	na
Campelo-Parada et al21	84.3% AVB9.3% BradycardiaOthers 6.2%	na	9.2%	Ventricular pacing >1% at 1 mo = AVB resolution	67.2%	1	na	• BAV associated with increased risk of conduction abnormalities persistence	na
Mirolo et al22	68.8% AVB30% others	2,5	9.3%	Ventricular pacing ≥ 1% = significant	75%	2.9	na	na	1.25% endocarditis lead leading to pacemaker explanation2.5% partial left pneumothorax secondary to subclavian vein puncture1.25% ventricular lead deplacement
Van Gils et al23	96% AVB5% SSS	5	9.3%	Less than 20% ventricular pacing over 6 mo' follow-up	25%	6	na	na	na
Takahashi et al24	na	na	16.4%	Absence, inadequate intrinsic ventricular rhythm, or ventricular pacing >95% in pacemaker interrogation during follow-up (PPM on VVI 30/min)	52.8%	13	na	• DDD mode and SE valves use associated with pacemaker dependency	na
Chan et al25	na	na	13.1%	Ventricular pacing reported	59%	12	na	na	1.6% atrial lead dislodgement6% ventricular lead dislodgement
Ghannam et al26	100% AVB	2,4	14%	No recovery of AV nodal conduction if CHB, high-grade AVB, or native ventricular rate <50 beats/min in absence of normal AV conduction	50%	12	na	• Larger aortic annulus less likely to recover conduction	1.2% (1 patient with right ventricular lead fracture)
Costa et al27	84.8% AVB4.1%SSS11% Others	na	13%	Absence of an escape or intrinsic rhythm for 30 s during temporary back-up pacing at a rate of 30 bpm	33.3%	12	na	• PPI associated with increased 6 y mortality • Baseline RBBB greater chance of being dependent at follow-up	na
Dolci et al28	80%AVB11% Bradycardia9% others	4	13%	“paced rhythm” reported	7%	12	• Baseline RBBB • QRS width immediately after TAVI	na	na
Tovia-Brodie et al29	92% AVB8% others	na	8.8%	No need for ventricular pacing defined as <1% ventricular pacing and intrinsic 1:1 AV conduction with the device programmed to VVI 30 beats per minute	39%	28,2	• Baseline long PR interval • Use of newer generation valves	na	3.7% tamponade
Junquera et al30	na	6	12.7%	AVB/CHB recovery = ventricular pacing rate <1%	33.4%	12	na	na	na
Meduri et al31	90% AVB6% Bradycardia4% others	2	28.4%	Patients who were symptomatic or did not have a native rhythm+capture of the percentage of paced ventricular beats	50%	12	• Baseline RBBB • Mean depth of valve implantation	• Medically treated diabetes mellitus in LOTUS valve patients	na
Total	82.7% AVB2.7% SSS2.8% bradycardia11.8% Others	3.2	17%	na	na	11.8	na	na	na

Values are n (%). PPI, Pacemaker implantation; AVB, atrioventricular block, na, not available; SSS, sick sinus syndrome; RBBB, right bundle branch block; BMI, body mass index; TAVI, transcatheter aortic valve implantation; SE, self-expandable; BAV, bicuspid aortic valve; PPM, permanent pacemaker; VVI, single-chamber device; DDD, dual-chamber device; AV, atrioventricular; CHB, complete heart block.

Follow-up is reported as mean or median as given by the authors.

Table E3

Pacemaker details in studies including < 250 patients (n = 30)

Study	Indications PPI	Timing of PPI, d	PPI rate	Dependency definition	Dependency rate	Dependency follow-up, months∗	Multivariable Predictors PPI	Association	PPI-related complications
Sinhal et alE1	na	2	5.6%	na	86%	6	na	na	na
Jilaihawi et alE2	90% AVB10% SSS	na	33.3%	Description by case of the % ventricular pacing at follow-up	66.6%	1	• LBBB with left axis deviation • Diastolic interventricular septa dimension >17 mm • Noncoronary cuspid thickness >8 mm	na	na
Baan et alE3	100% AVB	3	26.9%	“ventricular pacing” reported	100%	1	na	• Smaller left ventricle outflow tract diameter • More left-sided heart axis • More mitral annular calcification • Smaller postimplantation indexed effective orifice area	na
Fraccaro et alE4	96% AVB4% SSS	na	39%	With pacemaker to VVI at the lowest rate possible: continuous pacing or complete AVB or AF with inadequate ventricular response	23.5%	12	• Depth of the prosthesis implantation • Pre-existing RBBB	na	na
Van der boon et alE5	83.3% AVB13.8% bradycardia2.7% others	8	21.6%	By temporarily turning off the PPM or programming to a VVI modus at 30 bpm to assess dependency → if high-degree AVB (second degree Mobitz II or third degree) or a slow (<30 bpm) or absent ventricular escape rhythm observed	44.4%	11.5	na	na	na
Pereira et alE6	100% AVB	3	33%	Absence of any intrinsic or escape rhythm during back-up pacing at 30 beats/min (VVI)	27%	12	na	• Porcelain aorta = independent predictors of pacing dependency at follow-up	na
Goldenberg et alE7	61.5% AVB3% SSS34% others	na	16.8%	High degree of AVB (second degree or complete) or intrinsic rhythm <30 beats/min during pacemaker inhibition	29%	12	na	na	na
Ramazzina et alE8	46% AVB55% others	na	36.1%	>99% ventricular pacing	29%	12	• Use of MCV	• Diabetes • Atrial fibrillation before TAVI • Associated with no need for PPI	na
Boerlage-Van Dijk et alE9	91.3% AVB	3	19%	Ventricular-paced rhythm (no other definition)	52%	11.3	• Mitral annular calcification • Pre-existing RBBB	• No factors found	4.3% atrial lead repositioning4.3% pocket hematoma4.3% cerebral vascular accident
Renilla et alE10	100% AVB	2	37.9%	Presence of a high-degree AVB (Mobitz II and III) or a slow <30 bpm or absent ventricular escape rhythm (pacemaker turned off or programmed to VVI modus at 30 bpm)	39.1%	35	na	na	3% pacemaker pocket infection
Petronio et alE11	na	na	24.4%	VVI programming 30 beats/min	40.7%	1	• Implantation depth	• Implantation depth <4 mm	na
Weber et alE12	90% AVB2% bradycardy2% SSS5% others	na	23%	Pacemaker is partially or frequently needed to ensure heartbeat	35%	9	na	na	na
Schernthaner et alE13	78% AVB7% SSS13% Others	7	20%	Absence of an escape or intrinsic rhythm for 30 s during temporary back-up pacing at a rate of 30 bpm	37%	1.5	na	na	na
Kostopoulou et alE14	100% AVB	na	22%	Asystole or CHB with or without escape rhythm after cessation of pacing	9%	12	• Prolonged HV interval prognostic for PPI	• Trend between Δ QRS and PPI at 6-mo analysis	na
Sideris et alE15	100% AVB	na	38.7%	High ventricular pacing rate >99%	100%	6	na	na	1.5% pneumothorax
Luke et alE16	100% AVB	na	39.3%	“Pacing"	0.7%	3	• Previous conduction system disease • History of atrial fibrillation	Trends with• history of atrial fibrillation • presence of RBBB • male sex • atrioventricular nodal blocking drugs	na
Makki et alE17	63% AVB4% bradycardia33% others	3	14%	Underlying ventricular asystole >5 s, CHB, >50% pacing, symptoms in the setting of bradycardia (rate <50 bpm)	33%	3	na	na	na
Nijenhuis et alE18	87% AVB	8	24%	Ventricular pacing reported	68%	27	• Previous atrial fibrillation • Use of digoxin • MCV implantation • Left heart axis	na	na
Naveh et alE19	100% AVB	na	34.5%	Absent or inadequate intrinsic ventricular rhythm on pacemaker interrogation (intrinsic rhythm <30 bpm); >5% VP from the last follow-up on pacemaker interrogation; any evidence of VP on pacemaker interrogation in case where the programmed AV interval was >300 ms	68.4%	12	• Baseline RBBB • PR interval (each increment of 10 milliseconds in PR interval, risk for PPI 17% greater)	na	0%
Alasti et alE20	89.2% AVB2.6% SSS7.8% Others	3	25%	The need for ventricular pacing when the pacing rate was lowered to 40 bpm for 10 s → dependent: slow (<40 bpm) or absent ventricular escape rhythm or AVB (Mobitz II or III)	38%	12	na	na	16% hematomas
Ortak et alE21	83% AVB16% others	na	22%	na	64%	1	• Implantation depth • LOTUS implantation depth >4.8 mm = cut-off to predict PPI	na	na
Rodes-Cabau et alE22	81% AVB9% bradycardia9% others	42	11%	na	78%	12	na	na	na
Leong et alE23	74.6% AVB20.8% bradycardia13.5% others	2.3	44.8%	Ventricular pacing reported	73%	2.4	na	• Male sex • Increase in QRS duration post-TAVI • Associated with PPI	0%
Sharma et alE24	na	na	11.1%	Spontaneous response ventricular rate less than 30 bpm during backup pacing set at 30 beats/min for 30 s	32%	1	na	• RBBB • Intra-procedural CHB • Bifascicular block • QRS duration >120 ms • All associated with pacing dependence at 30 d	na
Bacik et alE25	49.8% AVB6.3% SSS43.9% others	5.5	13.8%	More than 95% pacing events	12%	12	• Weight • Absence of AF • Aortic peak gradient • Aortic valve area • Severity of pulmonary hypertension	na	na
Megaly et alE26	50.6% AVB34.9% others	na	35%	CHB requiring ventricular pacing	3.5%	12	na	na	na
Yazdchi et alE27	78% AVB	2	8.7%	Ventricular pacing reported	87%	14	na	na	na
McCaffrey et alE28	na	na	11.2%	Ventricular pacing reported	75%	1	na	• Predictive factors of acute conduction abnormalities (4) • Predictive factors of new conductions abnormalities after discharge (5)	na
Miura et alE29	90% AVB10% SSS	6	5%	Ventricular pacing reported	40%	12	na	na	na
Dhakal et alE30	80% AVB17% SSS3% others	2	17%	Ventricular pacing rate	54%	2.7	RBBB	In univariate analysis:• Valve size • SE valve • RBBB • Prolonged PR interval	na
Total	72% AVB1.7% SSS3% bradycardia23.2% others	6,5	23.8%	na	na	9	na	na	na

Values are n (%). PPI, Pacemaker implantation; na, not available; AVB, atrioventricular block; SSS, sick sinus syndrome; LBBB, left bundle branch block; VVI, single-chamber device; AF, atrial fibrillation; RBBB, right bundle branch block; PPM, permanent pacemaker; MCV, Medtronic CoreValve; TAVI, transcatheter aortic valve implantation; CHB, complete heart block; SE, self-expandable.

Follow-up is reported as mean or median as given by the authors.

Pacemaker Dependency

There was a large heterogeneity in the assessment and definition of the pacemaker dependency at follow-up. Indeed, 43% of the studies (10/23)10, 11, 12,,,18, 19, 20,, reported a right ventricular pacing rhythm as indicator of the pacemaker dependency at follow-up, whereas the remaining studies provided a wide range of pacemaker dependency definitions and evaluation.

Of the selected 23 articles, the majority (15 studies10, 11, 12, 13, 14, 15,,,24, 25, 26, 27, 28,,) reported the 1-year pacemaker dependency (Figure 2). The pacemaker dependency rate varied across time (Figure 3); it was 51.2% (16.7%-84.6%, n = 4) already at discharge, 57.9% (43.1%-71.3%, n = 10) at 1 month, 45.3%% (27.5%-64.5%, n = 8) at 6 months, and 49.5% (37.1%-61.9%, n = 15) at 1 year.

Figure 2

Pacemaker dependency at 1 year. Bars represent 1-year pacemaker dependency. The rate of pacemaker dependency ranged from 7% to 89% in individual studies.

Figure 3

Rate of pacemaker dependency across the time after TAVI. Pooled percentage is reported with 95% confidence limits (blue line). Light blue bars represent number of studies. Yellow line is the interpolation line. UCL, Upper confidence limit; LCL, lower confidence limit.

Influence of Baseline Right Bundle Branch Block (RBBB) and Atrial Fibrillation (AF) on 1-Year Pacemaker Dependency

In 6 studies,,,,,, among 531 patients undergoing pacemaker interrogation at 1 year, pacemaker dependency was present from 22% to 54% of the patients, and the difference between patients with and without baseline RBBB could be analyzed. The pooled OR was 1.76 (95% CI, 1.06-2.93). There was low heterogeneity (I2 = 14%) among the studies (Figure 4), nor was there publication bias (Figure E1). The Egger's test was not significant (P = .90). Leave-one-out study analysis (Table E4), as shown in the paper by Chan and colleagues influenced the pooled analysis, since leaving it in, the pooled estimates was not significant anymore. The association between preimplant RBBB and 1-year pacemaker dependency was not influenced by age (r = 0.0917, P = .6050), time of PPI (r = 0.0918, P = .9222), or sample size (r = 0.0920, P = .9245).

Figure 4

Impact of baseline RBBB on 1-year rate of pacemaker dependency Forest plot. Patients with baseline RBBB have 2-fold greater risk to develop pacemaker dependency 1 year after TAVI. TE, Log odds ratio; SE, standard error; IV, weighted mean difference; CI, confidence limits; PM, pacemaker.

Figure E1

Funnel plot on the impact of baseline RBBB on 1-year rate of pacemaker dependency. No publication bias was found among studies reporting the influence of baseline RBBB on pacemaker dependency.

Table E4

Leave-one-study out analysis

Left-out study	OR	95% CI	P value
RBBB
Kaplan 201915	2.3798	1.1797-4.8008	.0155
Nadeem 201820	2.2425	1.0932-4.6002	.0276
Raelson 201713	2.0960	1.0400-4.2241	.0385
Costa 201927	1.9345	1.0126-3.9288	.0479
Meduri 201931	1.8850	1.0136-3.7261	.0483
Chan 201825	1.6386	0.8001-3.3559	.1770
AF
Kaplan 201915	1.2220	0.4251-3.5122	.7098
Nadeem 201820	0.9829	0.3378-2.8599	.9748
Raelson 201713	0.9094	0.2919-2.8336	.8699
Chan 201825	1.4551	0.5188-4.0811	.4760
SE
Kaplan 201915	1.9233	1.0326-3.7266	.0426
Nadeem 201820	2.3432	1.1402-4.8153	.0205
Raelson 201713	2.5419	1.3676-4.7244	.0032
Van Gils 201712	2.3947	1.1342-5.0562	.0220
Urena 201410	1.8686	0.8743-3.9938	.1067
Takahashi 201824	1.8711	1.0048-3.4844	.0483

OR, Odds ratio; CI, confidence interval; RBBB, right bundle branch block; AF, atrial fibrillation; SE, self-expandable.

The impact of baseline AF could be investigated in 4 studies,,, and revealed no effect on 1-year pacemaker dependency (pooled OR, 1.71; 95% CI, 0.83-3.53) (Figure 5). Again, there was neither heterogeneity among the studies nor publication bias (Figure E2) The Egger's test was not significant (P = .79). On-leave out study analysis (Table E4) confirmed as AF had no impact on pooled analysis.

Figure 5

Impact of baseline AF on 1-year rate of pacemaker dependency. Forest plot. Patients with baseline AF have no greater risk to develop pacemaker dependency 1 year after TAVI. TE, Log odds ratio; SE, standard error; IV, weighted mean difference; CI, confidence limits; PM, pacemaker.

Figure E2

Funnel plot on the influence of baseline AF on pacemaker dependency. No publication bias was found among studies reporting the influence of baseline AF on pacemaker dependency.

Type of Implanted Prosthesis on 1-Year Pacemaker Dependency

The comparison between SE and BE protheses in terms of 1-year pacemaker dependency could be evaluated in 6studies (796 patients).,,,,, Patients who received SE prostheses had 2-fold greater risk for pacemaker dependency 1 year after TAVI (Figure 6). moderate heterogeneity was found (I2 = 43%). No publication bias was identified (Figure E3). The Egger's test was not significant (P = .5658). Leave-one-out study analysis (Table E4) showed as the paper by Urena and colleagues influenced the pooled analysis, since leaving it, the pooled estimates was not significant anymore. The association between preimplant SE and 1-year pacemaker dependency was not influenced by age (r = 0.1265, P = .5012), time of PPI (r = –1.1506, P = .0.6262), or sample size (r = 0.1123, P = .9812). The pacemaker dependency rate was significantly greater in those studies including more than 50% of SE prostheses (Figure E4).

Figure 6

Impact of SE prosthesis (vs BE) on 1-year rate of pacemaker dependency. Forest plot. Patients with SE prosthesis have 2-fold greater risk to develop pacemaker dependency at 1 year after TAVI. TE, Log odds ratio; SE, standard error; IV, weighted mean difference; CI, confidence limits; PM, pacemaker.

Figure E3

Funnel plot on the influence of SE versus BE valves on pacemaker dependency. No publication bias was found among studies reporting the influence of SE versus BE on pacemaker dependency.

Figure E4

Forest plot pooling pacemaker dependency according to percentage of SE prosthesis included in the study. IV, Weighted mean difference; CI, confidence interval; SE, self-expandable.

Third-Degree Atrioventricular Block and 1-Year Pacemaker Dependency

One-year pacemaker dependency rate was not significantly different between study where pacemaker was implanted due to third-degree atrioventricular block in a rate ranging from70%-79% (41.4%; 26.1%-58.5%), 80%-99% (48.3%; 21.3%-76.5%), and 100% (53.8%, 45.6%-61.7%), P = .427.

Complications and Multivariable Predictors of PPI

Only 5 studies,,,, reported the PPI-related complications. The rate of PPI-related complications ranged from 1.2% to 6%. The list of various complications is reported in Table 2. Information about predictors of PPI were provided by 7 studies (5319 patients). Pre-existing RBBB was the most frequent determinant factor of PPI. Regarding pacemaker dependency after TAVI at follow-up, multivariable analysis to investigate the predictors was performed in 5 studies: early PPI after TAVI, PPI on day 1, and larger aortic annular size, were found as independent predictor of persistent atrioventricular conduction disturbances. In contrast, 2 studies failed to demonstrate significant predictors for atrioventricular conduction recovery or pacemaker dependency.,

Discussion

The main findings of this systematic review regarding PPI and subsequent recovery of atrioventricular conduction or persistent pacemaker dependency in TAVI are as follows (Figure 7): (1) up to 50% of the patients with PPI after TAVI exhibit no pacemaker dependency at 1 year follow-up; (2) patients with baseline RBBB and (3) SE prosthesis have 2-fold greater risk to develop pacemaker dependency 1 year after TAVI; and (4) baseline AF does not influence the risk of pacemaker dependency at 1 year.

Figure 7

Up to 50% of the patients with permanent pacemaker implantation following TAVI exhibits no pacemaker dependency at 1-year follow-up. TAVI, Transcatheter aortic valve implantation; PPI, permanent pacemaker implantation; RBBB, right bundle branch block; TE, log odds ratio; SE, standard error; IV, weighted mean difference; CI, confidence limits; PM, pacemaker.

Several recent reports show that atrioventricular conduction defects requiring PPI post-TAVI may involve as much as 30% of the treated patients, therefore representing the most frequent complication in such a setting. However, nearly 50% of such patients are not pacemaker-dependent at 1-year follow-up, with recovery of atrioventricular conduction occurring even at a very early stage after implant, like before hospital discharge. Nonetheless, the range of postimplant pacemaker dependency or effective atrioventricular conduction restoration varies largely, ranging from 7% to 89%. This wide interval could be explained by heterogeneity and lack of consensus on pacemaker dependency definition. This limitation negatively influenced the study data interpretation in our analysis. Furthermore, pacemaker dependency may intermittently occur, thereby characterizing a different pattern of pacing dependency. This peculiar aspect was not available but might represent an additional factor to be investigated.

In our analysis, patients receiving SE valves and presenting with a preoperative RBBB had a 2-fold greater risk to have persistent pacing dependency at 1 year after TAVI. Ramazzina and colleagues (Table E2) showed that atrioventricular block as indication for PPI was always associated with pacemaker dependency at follow-up. In contrast, Gaede and colleagues demonstrated a low rate of long-term persistence of atrioventricular block after TAVI procedure. A few studies address investigate predictors of pacemaker dependency after TAVI. Naveh and colleagues (Table E2) showed that baseline RBBB, long post-TAVI PR interval, and delayed PR interval from baseline were independent predictors for long-term pacemaker dependency. Sharma and colleagues (Table E2) showed that bifascicular block, RBBB, intraprocedural complete heart block, and QRS duration >120 milliseconds were associated with pacemaker dependency at 30-day follow-up. The impact of baseline AF on pacemaker dependency at follow-up is controversial. Early PPI after TAVI procedure was the strongest predictor regarding persistent atrioventricular block and pacemaker dependency, based on large sample-size studies., According to these findings, we should emphasize that patients without baseline characteristics potentially leading to pacemaker dependency should benefit from other temporary leadless system as Micra AV (Medtronic, Minneapolis, Minn) to reduce the rate of permanent pacemaker implanted, as the Micra AV system is recently found to be safe; efficient and as performant as transvenous single-chamber pacemaker.,

Nowadays, guidelines regarding timing of PPI after TAVI are rather cloudy and not based on thorough clinical investigations., Due to the lack of consistent data, the dilemma about the appropriate timing for pacemaker implantation after TAVI is left to the discretion of the attending cardiologist according to the different centers' policies and therefore is associated with extreme variability in clinical management. Erkapic and colleagues showed that atrioventricular block occurs in more than 90% of the cases within the first post-TAVI week, which would allow to monitor carefully the patients at least seven days before considering PPI. However, some others studies support early PPI after TAVI procedure. Actually, as the optimal timing for PPI after TAVI is not established, the variability of decision certainly influences the outcome of pacemaker dependency, making difficult to conclude about the best interval to proceed to a definitive PPI.

PPI after TAVI is associated with increased long-term mortality. Faroux and colleagues show the negative impact of PPI after TAVI on survival and heart failure hospitalization within the year following TAVI. Xi and colleagues also show a greater all-cause mortality in TAVI patients receiving a PPI. In this systematic review, only 4 studies addressed the PPI-related complications and only 8 studies reported the 30-day mortality, making the overall appraisal of the impact of PPI on patient outcome likely underestimated, as emphasized by the Danish experience of Kirkfeldt and colleagues. Report on PPI-related complications is also markedly variable and the lack of a standardize international classification of these complications may further contribute to PPI-related complications and related outcome actually under-reported.

Limitations

This systematic review had several limitations that should be acknowledged. First, the follow-up varied remarkably among the included studies. The 1-year dependency was not available for all studies, limiting the conclusion to only 15 studies, whereas the others publications provided a follow-up mainly to 1 to 3 months. Only 3 studies presented a longer follow-up up to 4 years. Second, the heterogeneity about pacemaker dependency definition in the studies semantically limits our ability to compare the studies. Third, although several studies identified some risk-factors for PPI and only 1 study states multivariable predictors of pacemaker dependency, there was no agreement between studies. Finally, this is a systematic review of the literature; analysis of individual patients-data level may provide further understandings: in many of studies it was not specifically clear if the patient's native rhythm was assessed and thereby knowing for certain which patients truly required PPI.

Conclusions

This systematic review investigates the rate and predictors of pacing dependency after TAVI. Data from literature show that almost one half of the pacemakers are actively operating at 1-year follow-up. Baseline RBBB and SE valves are associated with a greater rate of pacemaker dependency at follow-up. These findings suggest that atrioventricular conduction disturbances after TAVI are reversible in a large percentage of patients. Such a condition may occur at variable time after the TAVI procedure, even within the TAVI-related hospitalization and, therefore, at a very early stage. These findings clearly indicate the need of thorough investigations regarding timing of pacemaker implantation, recovery of atrioventricular conduction, and predictors of pacemaker dependence as endpoints for further studies.

Conflict of Interest Statement

Dr Vernooy reported research grant from . Dr Van't Hof reported grants from Medtronic, Astra Zeneca, Boehringer Ingelheim, and Abbott. Dr Lorusso reported grants from Medtronic, LivaNova, and Eurosets. All other authors reported no conflicts of interest.

The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.