Pre-Existing Left Bundle Branch Block and Clinical Outcomes After Transcatheter Aortic Valve Replacement

Background Few reports on pre-existing left bundle branch block (LBBB) in patients undergoing transcatheter aortic valve replacement (TAVR) are currently available. Further, no present studies compare patients with new onset LBBB with those with pre-existing LBBB. Objectives This study aimed to investigate the association between pre-existing or new onset LBBB and clinical outcomes after TAVR. Methods Using data from the Japanese multicenter registry, 5,996 patients who underwent TAVR between October 2013 and December 2019 were included. Patients were classified into 3 groups: no LBBB, pre-existing LBBB, and new onset LBBB. The 2-year clinical outcomes were compared between 3 groups using Cox proportional hazards models and propensity score analysis to adjust the differences in baseline characteristics. Results Of 5,996 patients who underwent TAVR, 280 (4.6%) had pre-existing LBBB, while 1,658 (27.6%) experienced new onset LBBB. Compared with the no LBBB group, multivariable Cox regression analysis showed that pre-existing LBBB was associated not only with a higher 2-year all-cause (adjusted HR: 1.39; 95% CI: 1.06-1.82; P = 0.015) and cardiovascular (adjusted HR: 1.60; 95% CI: 1.04-2.48; P = 0.031) mortality, but also with higher all-cause (adjusted HR: 1.43, 95% CI: 1.07-1.91; P = 0.016) and cardiovascular (adjusted HR: 1.81, 95% CI:1.12-2.93; P = 0.014) mortality than the new onset LBBB group. Heart failure was the most common cause of cardiovascular death, with more heart failure deaths in the pre-existing LBBB group. Conclusions Pre-existing LBBB was independently associated with poor clinical outcomes, reflecting an increased risk of cardiovascular mortality after TAVR. Patients with pre-existing LBBB should be carefully monitored.

T ranscatheter aortic valve replacement (TAVR) is an established therapy for symptomatic severe aortic stenosis (AS). 1,2Preprocedural electrocardiography is particularly important to stratify the risks of patients undergoing TAVR.
Pre-existing right bundle branch block (RBBB) is well recognized as a risk factor for permanent pacemaker implantation (PPI) after the procedure and increases the risks of all-cause and cardiovascular mortality. 3,4However, the prognostic impact of preexisting left bundle branch block (LBBB) has not been well investigated.
6][7] The caveat is that these reports excluded patients with pre-existing LBBB, and only a few studies were conducted to investigate the impact of pre-existing LBBB on clinical outcomes.Fischer et al 8 reported that pre-existing LBBB was a risk for early PPI after TAVR, but not for late PPI, and had no significant impact on mortality after TAVR.Given that the prognostic impact could be different between preexisting and new onset LBBB, they should be separately treated in the analysis.Therefore, in the present study, we aimed to investigate the association with pre-existing or new onset LBBB and clinical outcomes in patients who underwent TAVR for severe AS.

METHODS
DATA SOURCE.We analyzed the data from the OCEAN-TAVI (Optimized transCathEter vAlvular iNtervention-transcatheter aortic valve implantation) registry. 9A total of 7,393 patients were enrolled in the OCEAN-TAVI registry between October 2013 and December 2019.The OCEAN-TAVI registry is a prospective, multicenter, observational registry of patients who underwent TAVR at 20 centers in Japan.
The OCEAN-TAVI registry was registered with the University Hospital Medical Information Network Clinical Trial Registry and accepted by the International Committee of Medical Journal Editors (UMIN000020423).All study participants provided informed consent, and the registry was approved by the ethics committees of all participating institutions.Patients were followed up annually at the participating institutions.The events were site-reported from the participating institutions.To ensure consistency, the database was regularly audited by the data committee members.
STUDY POPULATION.The study flow is presented in Figure 1.We excluded patients who had missing data for electrocardiographic records at either baseline or discharge (n ¼ 360), patients with complete RBBB (n ¼ 839), and patients with unknown native QRS due to ventricular pacing of a permanent pacemaker (n ¼ 225).As a result, 5,996 patients were included in the analyses.
OUTCOMES.The primary outcomes were 2-year allcause cardiovascular and noncardiovascular mortality after TAVR.The secondary outcomes were death from heart failure, sudden cardiac death (SCD) after TAVR, 30-day mortality, and in-hospital complications.All-cause mortality, cardiovascular mortality, and complications were defined based on the Valve Academic Research Consortium-2 criteria. 10SCD was defined as death occurring within 1 hour of symptom onset if witnessed, or within the past 24 hours if not witnessed, according to the World Health Organization definition.Patients with confirmed sudden death due to terminal disease or noncardiac causes were not considered to have experienced SCD. 11

ELECTROCARDIOGRAPHY AND ECHOCARDIOGRAPHY.
Twelve-lead electrocardiography and transthoracic echocardiography were performed at baseline, before hospital discharge, and at the annual followup.The diagnosis of intraventricular conduction disturbances was based on the American Heart Association/American College of Cardiology Foundation/Heart Rhythm Society recommendations for the standardization and interpretation of the electrocardiogram. 12As in previous reports, 3,4,8 incomplete RBBB and LBBB were considered normal.All transthoracic echocardiographic parameters were measured according to American Society of Echocardiography guidelines. 13,14he degree of    The cumulative incidences of all-cause, cardiovascular, and noncardiovascular mortality; death from heart failure; and SCD were calculated using the Kaplan-Meier method.Log-rank test and Cox regression analyses were performed.We checked the proportional hazard assumptions of each outcome using Schoenfeld residuals (Supplemental Figure 1).Univariable Cox regression analyses were performed for 2-year clinical outcomes.Thereafter, multivariable analyses were performed to examine variables that were independently associated with all-cause, cardiovascular, and noncardiovascular mortality, as well as death from heart failure and SCD.In multivariable analysis, variables associated with mortality, based on previous studies, were included.[16][17][18][19][20][21][22]       To address the issue of residual confounding, sensitivity analyses using an array approach were performed to investigate how the observed HR may be affected by certain variations in the assumptions about the presence of an unmeasured confounder.25 This method is a statistical technique used to adjust for confounding factors that are difficult to adjust for or measure due to the nature of observational studies.26 The observed HRs for pre-existing LBBB using new onset LBBB as a reference were 1.43 for allcause mortality and 1.81 for cardiovascular mortality.The variables included in the multivariate Cox regression analyses for death from heart failure were following: age, sex, NYHA functional class III or IV, diabetes mellitus, chronic kidney disease, atrial fibrillation, previous stroke, peripheral artery disease, previous coronary artery bypass grafting, permanent pacemaker, B-type natriuretic peptide $400 pg/mL or N-terminal pro-B-type natriuretic peptide $1,600 pg/mL, left ventricular ejection fraction #40%, tricuspid regurgitation $moderate, myocardial infarction, acute kidney injury, bleeding, and paravalvular leak $moderate.The variables included in the multivariate Cox regression analyses for sudden cardiac death were following: NYHA functional class III or IV, albumin <3.5 g/dL, and bleeding.The variables included in the IPTW analyses were following: age, sex, body mass index, body surface area, NYHA functional class III or IV, hypertension, dyslipidemia, diabetes mellitus, chronic kidney disease, previous stroke, chronic obstructive pulmonary disease, peripheral artery disease, liver disease, coronary artery disease, previous coronary artery bypass grafting, atrial fibrillation, permanent pacemaker, clinical frailty score, beta-blockers, renin-angiotensin system inhibitors, nontransfemoral approach, local anesthesia, hemoglobin, estimated glomerular filtration rate, albumin <3.5 g/dL, B-type natriuretic peptide $400 pg/mL or N-terminal pro-B-type natriuretic peptide $1,600 pg/mL, left ventricular ejection fraction, aortic regurgitation $moderate, mitral regurgitation $moderate, tricuspid regurgitation $moderate, balloon expandable valve, stroke, myocardial infarction, vascular complications, acute kidney injury, bleeding, new pacemaker implantation, transcatheter heart valve mean pressure gradient, indexed effective orifice area, and paravalvular leak $moderate.
Abbreviations as in Table 2.
IPTW analyses were performed.The maximum absolute value of the standardized mean difference was <0.1 in all examined covariates in the weighted cohort (Supplemental Table 1).The number of included participants per group is shown Table 2.The comparison of participants included and excluded from IPTW analyses is shown in Supplemental  2).In the IPTW analysis, there was no significant difference in noncardiac death between the 3 groups.in SCD between the 3 groups (Table 3).The full univariable and multivariable model results are shown in Supplemental Tables 7 and 8.
SUBGROUP ANALYSES.Subgroup analyses for allcause mortality are shown in Supplemental Figure 3.
There were no significant interactions between preexisting LBBB and prespecified subgroups.
SENSITIVITY ANALYSES.Sensitivity analyses showed a curved surface representing the fully adjusted HRs for pre-existing LBBB using new onset LBBB as reference (Supplemental Figure 4).The prevalence of an unmeasured confounder in the new onset LBBB group (P C0 ) was set at 0.5.The prevalence of the unmeasured confounder in the pre-existing LBBB group (P C1 ) varied between 0.0 and 1.0 on the x-axis.The strength of the confounder-outcome association (HR CD ) varied between 1.0 and 5.5 on the z-axis.In allcause and cardiovascular mortality, all adjusted HRs were >1 in different settings.
POST HOC ANALYSES.The results of the post hoc analyses are shown in Supplemental Table 9.The results were similar to those of the main analyses.

DISCUSSION
The impact of pre-existing LBBB on patients undergoing TAVR remains unknown.The main findings of our study showed that pre-existing LBBB was associated with poor clinical outcomes after TAVR.
Furthermore, patients with pre-existing LBBB had worse prognoses than those with new onset LBBB.To the best of our knowledge, the present study is the first to identify the association of pre-existing LBBB and a high risk of mortality after TAVR.
Most reports of LBBB in patients who underwent TAVR pertain to postprocedural new onset LBBB.6][7] However, these reports exclude patients with pre-existing LBBB.As such, few studies are available on the impact of pre-existing LBBB on clinical outcomes after TAVR.Fischer et al 8 reported that preexisting LBBB is a risk for early PPI after TAVR, but not for late PPI, and, furthermore, has no significant effect on mortality after TAVR.However, in our study, pre-existing LBBB was not a risk for early PPI.
However, the patients with pre-existing LBBB had a higher rate of prior pacemaker implantation, precluding post-TAVR pacemaker implantation.In terms of complications, bleeding and pacemaker implantation were more common in patients with new onset LBBB.However, complications were similar for patients with pre-existing LBBB and no LBBB.
In our study, 4.6% of patients undergoing TAVR had LBBB.This percentage is greater than that of LBBB in the general population. 27However, the percentage of pre-existing LBBB patients undergoing TAVR reported so far is around 10%, 8  Based on the findings of prior studies, potential mechanisms of our results may include progression to high-degree AVB, ventricular arrhythmia, or worsening heart failure.However, in our study, many deaths were due to heart failure, not SCD.
In our study, patients with pre-existing LBBB had worse prognoses than those with new onset LBBB.
We also performed sensitivity analyses on the HR of Further investigation is required.
STUDY LIMITATIONS.First, as shown in Supplemental Table 2, patients not included in the study may have a poor prognosis, leading to selection bias.This may be related to the fact that patients with RBBB have a poor prognosis, 3,4 and those without electrocardiography recordings are in poor condition.Second, no detailed electrocardiographic information, such as PR interval and QRS duration, is available.QRS duration is important, as it has been reported to be associated with prognosis. 31In addition to new onset LBBB, there are also reports that first degree of atrioventricular block is a risk for PPI. 32Third, the indication of PPI during follow-up was not collected although the frequency of PPI during follow-up is important.Fourth, follow-up medication and echocardiographic data are scarce.There are reports that beta-blockers and renin-angiotensin inhibitors were associated with lower mortality in patients who underwent TAVR. 33,34The optimal medical therapy may be important because patients with pre-existing LBBB have lower ejection fraction than those with no LBBB.
Fifth, there is a lack of procedural and computed tomography data.The length of the membrane septum and the depth of valve implantation are relevant for new onset LBBB and permanent pacemaker implantation. 35,36Sixth, although the multivariate Cox regression and IPTW analyses were performed, the baselines of each group differed greatly.This study was an observational study and could only be adjusted with measured covariates.Therefore, the selection bias was inevitable and baseline characteristics were not completely aligned.Finally, although 3-group comparisons were used in this study, multiple comparisons are prone to statistical errors.
Therefore, more meticulous studies that consider these limitations should be our future targets.

CONCLUSIONS
Pre-existing LBBB was independently associated with poor outcomes, reflecting increased risk of cardiovascular mortality after TAVR.Patients with preexisting LBBB should be carefully monitored after TAVR.Further investigation will be required to corroborate our findings.
cardiac death TAVR = transcatheter aortic valve replacement University Hospital, Toyama, Japan; p Department of Cardiology, National Cerebral and Cardiovascular Center, Osaka, Japan; q Department of Cardiology, Teikyo University School of Medicine, Tokyo, Japan; r Department of Cardiology, Toyohashi Heart paravalvular leak (PVL) was measured in accordance with the guidelines and reported as a semiquantitative grade: none, trace, mild, moderate, or severe.15Pre-existing LBBB was defined as the presence of complete LBBB on the electrocardiography at baseline.New onset LBBB was defined the presence of complete LBBB on postprocedural electrocardiography before hospital discharge.STATISTICAL ANALYSIS.At first, baseline and 30-day outcomes and complications were compared between the 3 groups: pre-existing LBBB, new onset LBBB, and no LBBB (

FIGURE 1
FIGURE 1 Study Flowchart Pre-Existing LBBB in Patients Undergoing TAVR CENTRAL ILLUSTRATION All-Cause, Cardiovascular, and Noncardiovascular Mortality and Causes of Cardiovascular Death 0.0 Pre-existing vs no LBBB Pre-existing vs new onset LBBB New onset vs no LBBB Adjusted HR: 1.39 (95% CI: 1.06-1.82)Adjusted HR: 1.43 (95% CI: 1.07-1.91)Adjusted HR: 0.97 (95% CI: 0.83-1.13) Prevalence of the unmeasured confounder in new onset LBBB group was set at 0.5.Two factors were varied: the strength of the unmeasured confounderoutcome association (1.0-5.5) and the prevalence of the unmeasured confounder in the pre-existing LBBB group (0.0-1.0).A 3-dimensional mesh plot was constructed to check the impacts on the fully adjusted HR by varied unmeasured confounder settings.Finally, the post hoc analysis with the 2-arm comparison was added to confirm the consistency of the results found in the 3-arm comparison.Multivariate Cox regression and IPTW analyses were performed for all-cause mortality in each of the 2 groups.The loss of cases due to missing values in the multivariable analyses and propensity score analysis was 2.3%.The analysis of the pattern of missing values is shown in Supplemental Figure 2. Because the proportion of missing values was small, multiple imputation was not performed.All statistical analyses were performed using R software version 3.6.1 (R Foundation for Statistical Computing).All tests were 2-sided, and P values of <0.05 were considered statistically significant.RESULTS BASELINE CHARACTERISTICS AND 30-DAY OUTCOMES.

FIGURE 2
FIGURE 2 All-Cause, Cardiovascular, and Noncardiovascular Mortality in the Weighted Cohort

Figure 3 .
Figure 3.During the follow-up period, 173 patients died of heart failure, 43 of whom died of SCD.The

FIGURE 3
FIGURE 3 Death From Heart Failure and Sudden Cardiac Death
The IPTW ¼ inverse probability of treatment weighting; LBBB ¼ left bundle branch block.

TABLE 3
Analyses of Death From Heart Failure and Sudden Cardiac Death

Table 3 .
Heart failure death and the incidence in our study is 30 addition, new onset LBBB is caused by mechanical compression of the conduction system by a prosthetic valve, but pre-existing LBBB is due to a failure of one's own conduction system, which may be hiding cardiomyopathy or other problems that were not fully understood in our study, which may have led to our results.It has been reported that bundle branch block is more common in patients with amyloidosis,30and since our study excluded patients with RBBB, it is possible that there were more patients with amyloidosis in the pre-existing group.There is no available evidence for treatment of pre-existing or new onset LBBB after TAVR.Thus, further research is needed to determine the timing of device implantation and which devices, including cardiac resynchronization therapy, are best.Because the mortality rate is higher in patients with preexisting LBBB, it may be important to know whether patients with LBBB after TAVR have preexisting LBBB or new onset LBBB.Although perioperative risks are not high, patients with preexisting LBBB should be carefully monitored after TAVR.However, currently, there are no reports other than our study showing similar results.