The Prognostic Impact of New‐Onset Persistent Left Bundle Branch Block Following Transcatheter Aortic Valve Implantation: A Meta‐analysis

ABSTRACT

New‐onset persistent left bundle branch block (NOP‐LBBB) is one of the most common conduction disturbances after transcatheter aortic valve implantation (TAVI). We hypothesized that NOP‐LBBB may have a clinically negative impact after TAVI. To find out, we conducted a systematic literature search of the MEDLINE/PubMed and Embase databases. Observational studies that reported clinical outcomes of NOP‐LBBB patients after TAVI were included. The random‐effects model was used to combine odds ratios, risk ratios, or hazard ratios (HRs) with 95% confidence intervals. Adjusted HRs were utilized over unadjusted HRs or risk ratios when available. A total of 4049 patients (807 and 3242 patients with and without NOP‐LBBB, respectively) were included. Perioperative (in‐hospital or 30‐day) and midterm all‐cause mortality and midterm cardiovascular mortality were comparable between the groups. The NOP‐LBBB patients experienced a higher rate of permanent pacemaker implantation (HR: 2.09, 95% confidence interval: 1.12‐3.90, P = 0.021, I ² = 83%) during midterm follow‐up. We found that NOP‐LBBB after TAVI resulted in higher permanent pacemaker implantation but did not negatively affect the midterm prognosis. Therefore, careful observation during the follow‐up is required.

Introduction

Transcatheter aortic valve implantation (TAVI) has shown a comparable long‐term prognostic benefit compared with surgical aortic valve replacement (SAVR) for severe symptomatic aortic stenosis in both high and intermediate surgical risk.1, 2 However, several perioperative complications may diminish the benefit of TAVI and be observed more frequently than following SAVR. New‐onset persistent left bundle branch block (NOP‐LBBB) is one of the most frequent conduction disturbances observed following TAVI and could potentially attenuate the benefit of the procedure.3, 4

Left bundle branch block causes electrical and mechanical dyssynchrony of the heart and various unfavorable hemodynamic effects.5 Previous studies have reported its negative prognostic impact on wide variety of cardiac disorders.6, 7, 8 Hoffman et al have reported that new‐onset LBBB or permanent pacemaker implantation (PPI) following TAVI was an independent risk factor for reduced left ventricular ejection fraction (LVEF) at 12‐month follow‐up.9 More importantly, according to Urena et al, NOP‐LBBB is associated with increased sudden cardiac death following TAVI.10 However, the prognostic impact of NOP‐LBBB following TAVI has not yet been established. Schymik et al have reported NOP‐LBBB as an independent risk factor for all‐cause mortality, but other studies have not found an association between NOP‐LBBB and all‐cause mortality.11, 12, 13 Because new‐onset LBBB is a rare conduction disturbance with no clear association with mortality following SAVR, it is paramount to assess the prognostic impact of NOP‐LBBB following TAVI to improve its outcomes.14

Therefore, we sought to summarize the prognostic impact of NOP‐LBBB following TAVI through a meta‐analysis.

Methods

Search Strategy

A systematic literature search was conducted through February 2016. We set neither time nor language restrictions. Two independent reviewers (T.A. and H.T.) searched the MEDLINE/PubMed and Embase databases. Search terms: (1) “left bundle branch block” or “LBBB” or “conduction”; and (2) “transcatheter aortic valve” or “percutaneous aortic valve” or “TAVI” or “TAVR” or “transfemoral” or “transapical” or “transsubclavian” or “transaortic” or “transcarotid” or “Sapien” or “CoreValve” or “balloon expandable” or “self expandable.” We utilized a 2‐level search strategy. Titles or abstracts were screened based on inclusion and exclusion criteria. When considered relevant, full manuscripts were reviewed. The references of manuscripts included for full review were manually reviewed. When data were considered to have an overlap, only the most recent study was included. A systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) and Meta‐analysis of Observational Studies in Epidemiology (MOOSE) guidelines.15, 16

Inclusion and Exclusion Criteria

Studies were included if they met the following criteria: (1) comparative study design; (2) the study population comprised patients with and without NOP‐LBBB (defined as postprocedural onset and persistent at discharge) following TAVI for severe aortic stenosis; and (3) outcomes included perioperative (in‐hospital or 30‐day) all‐cause mortality and midterm (≥1‐year) all‐cause and cardiovascular mortality and PPI. Exclusion criteria were (1) single‐arm study of NOP‐LBBB patients with no comparison of groups; and (2) no clear definition as to whether LBBB was persistent or transient.

Data Extraction

Data from the included studies were filled into preformed Excel spreadsheets. These data included author(s), publication year, study location and design, baseline characteristics of patients, and outcomes of interest. We abstracted odds ratios, risk ratios (RRs), and hazard ratios (HRs) with their 95% confidence intervals (CIs) of perioperative and midterm outcomes of interest, respectively, for NOP‐LBBB vs non–NOP‐LBBB patients. When an HR was not described in the manuscript, it was calculated by using the methods reported by Tierney et al,17 if possible. If it was impossible to calculate an HR, a risk ratio was extracted instead of an HR. When adjusted RRs/HRs were available, they were preferentially abstracted and combined in a meta‐analysis. Data extraction from the studies was based on consensus between the 2 reviewers.

Outcomes

A primary outcome was midterm (≥1‐year) all‐cause mortality. Secondary outcomes were perioperative (in‐hospital or 30‐day) all‐cause mortality, midterm (≥1‐year) cardiovascular mortality, and need for PPI.

Statistical Analysis

Meta‐analyses were performed with Review Manager (RevMan) version 5.3 (Nordic Cochrane Centre, The Cochrane Collaboration, 2012, Copenhagen, Denmark) and Comprehensive Meta‐Analysis version 2 (Biostat, Englewood, NJ). The estimated pooled RR/HR was calculated by inverse variance weighted average of logarithmic RRs/HRs in the random‐effects model. Heterogeneity of the studies was assessed with the I ² index, which indicates 25%, 50%, and 75% as low, moderate, and high heterogeneity, respectively. Sensitivity analysis was performed by eliminating each study one at a time to assess its effect on the results. Significant heterogeneity was considered to be present when the P value was <0.05. Publication bias was evaluated visually by assessing the funnel plot and quantitatively with the Egger test when ≥10 studies were included in a meta‐analysis (http://handbook.cochrane.org); however, publication bias was not assessed, as <10 studies were included. A P value of <0.05 was considered significant.

Results

Search Result

We identified 7 comparative studies3, 4, 11, 12, 13, 18, 19 including a total of 4049 patients (807 patients with NOP‐LBBB and 3242 patients without NOP‐LBBB). The diagnosis criteria for LBBB were based on the recommendations from the American Heart Association/American College of Cardiology Foundation/Heart Rhythm Society20 in 4 studies3, 4, 13, 18 and from the World Health Organizational/International Society and Federation for Cardiology21 in 2 studies.11, 19 One study12 did not detail the diagnosis criteria. The study‐selection process is shown in Figure 1. Details of included studies are summarized in the Table 1.

Figure 1.

Flow diagram for study selection.

Table 1.

Basic Information on Included Studies and Patient Characteristics

Author	Year	Study Design	NOP‐LBBB Definition	NOP‐LBBB Cohort	Non–NOP‐LBBB Cohort	Follow‐up
Nazif	2014	Prospective, PARTNER trial	New LBBB present on discharge or 7‐day ECG	121	1030	12 months
Carrabba	2015	Prospective, single‐center	New LBBB that persisted at discharge	34	58	12 months
Houthuizen	2014	Prospective, multicenter	New LBBB that persisted at 1 year	111	365	Median 898 and 944 days for non–NOP‐LBBB (either did not develop new‐onset LBBB or transient LBBB) and 914 days for NOP‐LBBB
Urena	2014	Prospective, multicenter	New LBBB that persisted at discharge	79	589	Median 13 months (IQR, 3–27 months)
Schymik	2015	Prospective, Karlsruhe registry	New LBBB that persisted at discharge	197	437	12 months
Franzoni	2013	Unclear, single‐center	New LBBB that persisted at discharge	41	197	12 months
Testa	2013	Prospective, multicenter	New LBBB that persisted at discharge	224	594	12 months

Author	Patient Age	Male Sex, %	HTN, %	DM, %	CAD, %	PAD, %	CKD, %	Logistic EuroSCORE	TF Access, %	Valve
Nazif	84.2 ± 7.2	43.7	92.6	37.5	37.7 (CABG), 38.1 (PCI)	43.0	15.8 (Cr ≥2.0 mg/dL)	24.7 ± 15.6b	56.6	ESV 100%
Carrabba	81 ± 6.3	52	60.0	30	41.3 (CABG and PCI)	9.8	18.5	20 ± 14	97	MCRS 100%
Houthuizen	Median 81 (IQR, 77–85)	43.7	NR	26.9	59.7 (CABG and PCI)	25.6	1.10a (IQR, 0.86–1.41)	16.4 (IQR, 10.1–25.4)	63.2	ESV 53.2%, MCRS 46.8%
Urena	80.7 ± 8.2	48.7	81.4	30.5	33.1 (CABG)	NR	43.9	21.2 ± 14.1	54.3	ESV 100%
Schymik	82.02 ± 4.46	38.3	NR	33.8	15.1 (CABG)	12.9	5.8 (Cr ≥2.2 mg/dL)	21.71 ± 13.14	NR	ESV 80.8%, MCRS 19.2%
Franzoni	79.4 ± 7.6	53.8	75.2	27.3	39.1% (CAD)	NR	55.5	22.4 ± 15.1	NR	ESV 63.4%, MCRS 36.6%
Testa	82 ± 6.5	45.5	NR	24.4	30.2% (previous PCI)	19.6	NR	23.41 ± 11.3	88.3	MCRS 100%

Abbreviations: CABG, coronary artery bypass grafting; CAD, coronary artery disease; CKD, chronic kidney disease; Cr, creatinine; DM, diabetes mellitus; ECG, electrocardiogram; ESV, Edwards Sapien valve; EuroSCORE, European System for Cardiac Operative Risk Evaluation; HTN, hypertension; IQR, interquartile range; LBBB, left bundle branch block; MCRS, Medtronic CoreValve Revalving System; NOP‐LBBB, new‐onset persistent left bundle branch block; NR, not reported; PAD, peripheral artery disease; PARTNER, Placement of Aortic Transcatheter Valves; PCI, percutaneous coronary intervention; SD, standard deviation; TF, transfemoral.

Continuous numbers are expressed as either mean ± SD or median (IQR).

^aCr (mg/dL).

^bEuroSCORE.

Meta‐analysis of Outcomes

The midterm (≥1‐year) all‐cause mortality rate (807 and 3242 patients with and without NOP‐LBBB, respectively) was 24.4% in the NOP‐LBBB group and 20.9% in the non–NOP‐LBBB group. There was no significant difference in midterm all‐cause mortality between the NOP‐LBBB and non–NOP‐LBBB groups (HR: 1.13, 95% CI: 0.89‐1.44, P = 0.31, I ² = 59.7%; Figure 2). There was significant heterogeneity observed (P = 0.021). Sensitivity analyses did not significantly alter the result.

Figure 2.

Forest plot for midterm (≥1‐year) all‐cause mortality in the NOP‐LBBB and non–NOP‐LBBB groups. Abbreviations: CI, confidence interval; NOP‐LBBB, new‐onset persistent left bundle branch block.

Perioperative all‐cause mortality (542 and 2061 patients with and without NOP‐LBBB, respectively) was 5.5% in the NOP‐LBBB group and 3.8% in the non–NOP‐LBBB group. There was no significant difference in perioperative all‐cause mortality between the NOP‐LBBB and non–NOP‐LBBB groups (odds ratio: 1.48, 95% CI: 0.82‐2.66, P = 0.20, I ² = 36.6%; Figure 3A). No significant heterogeneity was observed, and subsequent sensitivity analyses demonstrated similar results.

Figure 3.

Forest plots for (A) perioperative mortality and (B) midterm cardiovascular mortality in the NOP‐LBBB and non–NOP‐LBBB groups. Abbreviations: CI, confidence interval; NOP‐LBBB, new‐onset persistent left bundle branch block.

The midterm cardiovascular mortality rate (465 and 2410 patients with NOP‐LBBB and without NOP‐LBBB, respectively) was 9.2% in the NOP‐LBBB group and 7.7% in the non–NOP‐LBBB group. There was no significant difference in midterm cardiovascular mortality between the NOP‐LBBB and non–NOP‐LBBB groups (HR: 1.04, 95% CI: 0.81‐1.32, P = 0.79, I ² = 6.0; Figure 3B). No significant heterogeneity was observed, and subsequent sensitivity analyses indicated similar results.

The NOP‐LBBB patients (n = 621) required more frequent PPI than did the non–NOP‐LBBB patients (n = 2650) during midterm follow‐up (HR: 2.09, 95% CI: 1.12‐3.90, P = 0.021, I ² = 83%; Figure 4). There was significant heterogeneity (P = 0.001). The sensitivity analysis showed an insignificant result when studies by Nazif, Urena, or Schymik10, 11, 13 were removed.

Figure 4.

Forest plot for repeat hospitalization during midterm follow‐up in the NOP‐LBBB and non–NOP‐LBBB groups. Abbreviations: CI, confidence interval; NOP‐LBBB, new‐onset persistent left bundle branch block.

Discussion

The main findings of our meta‐analysis are that NOP‐LBBB patients experience PPI more frequently during midterm follow‐up, and NOP‐LBBB after TAVI was not associated with perioperative and midterm all‐cause mortality, as well as midterm cardiovascular mortality.

Permanent Pacemaker Implantation in New‐Onset Persistent Left Bundle Branch Block Cohort

The NOP‐LBBB patients had a higher PPI rate during the follow‐up. High heterogeneity was observed (I ² = 83%, P = 0.001). Self‐expandable balloon was a risk of both NOP‐LBBB and PPI,11, 18 and we consider that the significant heterogeneity is mainly because the implanted prosthesis differs considerably among studies, as summarized in the Table 1. The main risks for PPI after TAVI include preexistent right bundle branch block, self‐expandable prosthesis valve, and depth of implantation.22, 23 New‐onset LBBB was a predictive factor of high‐grade atrioventricular (AV) block following TAVI.24 Some reports showed that complete AV block, Mobitz type second‐degree AV block, and sinus‐node dysfunction were the common indications for PPI following TAVI.25 These data are, however, scarce for NOP‐LBBB patients who require PPI after TAVI. The main indications for PPI during the follow‐up in NOP‐LBBB patients in the present meta‐analysis were in agreement with those in past reports.25 The cause of the higher PPI rate in NOP‐LBBB patients is likely multifactorial. Patients who developed NOP‐LBBB may represent those with underlying conduction vulnerability, and therefore they may be susceptible to developing further conduction abnormalities during the midterm follow‐up.

Meta‐analysis was not performed for hospitalization in patients who developed NOP‐LBBB after TAVI because of insufficient data. The NOP‐LBBB patients had a higher hospitalization rate in 2 studies included in our meta‐analysis, although this was not statistically significant.10, 13 Patients with new‐onset LBBB and left‐axis deviation after SAVR suffer adverse events more frequently.26 Wider QRS duration in patients with NOP‐LBBB following TAVI was an independent risk of all‐cause mortality or admission due to heart failure (HF).27 Cardiac resynchronization therapy (CRT) has shown to be effective in reducing adverse events in LBBB patients.28 The benefit of CRT in NOP‐LBBB after TAVI is widely unknown. Meguro et al reported an effective CRT therapy for a patient with worse cardiac function with NOP‐LBBB after TAVI.29 Antonio et al reported duration of symptoms for <1 year to be an independent predictor of patients who respond better to CRT therapy.30 Because the nature of NOP‐LBBB is often unpredictable,31 further investigation to identify prolonged NOP‐LBBB is warranted in the future. Nombela‐Franco et al have reported readmission causes and rates in post‐TAVI patients. Nearly half of the patients experienced unplanned readmission; more than half (59%) of readmissions were due to noncardiac reasons and the rest were classified as cardiac. The main causes of noncardiac admission were respiratory diseases, infection, and bleeding.32 One possible explanation for increased admission in NOP‐LBBB patients is that these patients may be more vulnerable to HF exacerbation in the middle of these events (infection and bleeding). However, Urena et al and Testa et al reported no association between NOP‐LBBB and HF admission.13, 19 This may be due to the uncertain permanency of NOP‐LBBB.

New‐Onset Persistent Left Bundle Branch Block and Mortality

Perioperative and midterm all‐cause mortality, as well as midterm cardiovascular mortality, did not differ between the NOP‐LBBB and non–NOP‐LBBB groups. Right bundle branch block as well as LBBB have been associated with increased mortality.33, 34

Urena et al have reported that NOP‐LBBB is associated with decreased LVEF and poorer functional status at 1‐year follow‐up.13 In TAVI patients, lower LVEF is a negative prognostic factor.35 Also, patients with poor functional status have a 2‐fold increase of death following TAVI.36 Furthermore, LBBB is associated with increased cardiovascular mortality in a half of patients with chronic HF.37 However, results of our meta‐analysis proved the association of NOP‐LBBB with neither all‐cause nor cardiovascular mortality. There are several potential reasons for our observation. First, we included studies reporting outcomes of “persistent” (continuing until discharge or ≥7 days) new‐onset LBBB. The natural course of NOP‐LBBB, however, is unclear. Nazif et al have reported a decrease in proportion of patients with LBBB from 10.5% at discharge to 8.5% at 6 to 12 months.3 Testa et al reported similar LBBB rates at discharge and 30 days after TAVI.19 The lack of data regarding the natural course of NOP‐LBBB may be one of the reasons for inconsistent results of mortality. Results of the study by Houthuizen et al showed a significant increase in mortality in the NOP‐LBBB group, when NOP‐LBBB was defined as LBBB persisting during 1‐year follow‐up.18 Furthermore, TAVI patients with underlying chronic LBBB have increased 1‐year mortality.38 This finding may also support the idea that unclear persistency of NOP‐LBBB may be one of the reasons for inconsistency of mortality in NOP‐LBBB following TAVI. However, limited data suggest that a high proportion of NOP‐LBBB patients will remain in LBBB.3, 19 Second, most of the studies reported 1‐year (not beyond) mortality. The study by Houthizen et al had the longest follow‐up duration among included studies and revealed increased mortality in NOP‐LBBB patients. A study reported that the mean duration to develop HF in LBBB patients was 3.3 years.39 It would be interesting to know if longer duration of follow‐up will result in different mortality rates. Third, because of the uncertainty of the NOP‐LBBB course, PPI often may be overused in these patients. Ramazzina et al reported that out of 17 patients who had PPI for new‐onset LBBB, none of them required >1% of ventricular pacing; whereas of those who required PPI for atrioventricular conduction disorder, 83% (10/12) required >1% of ventricular pacing based on pacemaker interrogation.40 This may suggest that PPI is overused in LBBB after TAVI and these patients may not benefit from PPI. Last, no studies have clearly reported whether the LBBB was strict or nonstrict. It was reported that strict LBBB patients had greater mechanical dyssynchrony compared with nonstrict LBBB patients.41 However, Sundh et al have reported that the majority of the new‐onset LBBB following TAVI met the strict LBBB criteria, and therefore we consider that the difference between strict and nonstrict LBBB criteria would have minimal effects on our outcomes.42

Study Limitations

This study possesses all the inherent bias associated with meta‐analysis. First, this was a meta‐analysis of observational studies and therefore susceptible to bias. Secondly, we abstracted and then combined “unadjusted” RR/HR when “adjusted” ones were unavailable. However, excluding “unadjusted” RR/HR could lead to greater bias in favor of NOP‐LBBB. When an “unadjusted” RR/HR is not statistically significant, often it may not be incorporated into a multivariate analysis and not reported in the manuscript. Third, although we mounted a rigorous effort in search of published manuscripts, we cannot totally exclude the possibility of publication bias. Fourth, most of the studies included had a follow‐up period of only 1 year. Finally, the meta‐analysis for perioperative mortality only included a limited number of studies (total of 3) and the pooled‐effect size was associated with a wide 95% CI. Therefore, the result must be interpreted with caution. However, there was no significant heterogeneity observed, and subsequent sensitivity analysis showed the same result. This implies that the result is robust.

Conclusion

New‐onset persistent left bundle branch block after TAVI was associated with an increased rate of PPI but did not negatively affect midterm all‐cause and cardiovascular mortality. Careful monitoring is required for indication of PPI during follow‐up. Longer follow‐up is warranted to further assess the impact of NOP‐LBBB.