Author + information
- Received August 13, 2018
- Revision received October 25, 2018
- Accepted October 30, 2018
- Published online January 7, 2019.
- Troels H. Jørgensen, MDa,∗ (, )
- Ole De Backer, MD, PhDa,
- Thomas A. Gerds, DrRerNatb,
- Gintautas Bieliauskas, MDa,
- Jesper H. Svendsen, MD, DMSca,c and
- Lars Søndergaard, MD, DMSca,c
- aDepartment of Cardiology, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark
- bSection of Biostatistics, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark
- cDepartment of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark
- ↵∗Address for correspondence:
Dr. Troels H. Jørgensen, Rigshospitalet, The Heart Centre, Blegdamsvej 9, 2100 Copenhagen, Denmark.
Objectives The aim of this study was to assess mortality and rehospitalization in patients with new bundle branch block (BBB) and/or permanent pacemaker (PPM) after transcatheter aortic valve replacement (TAVR).
Background Previous studies have provided inconsistent results on the clinical impact of new BBB or new PPM after TAVR.
Methods A total of 816 consecutive patients without pre-procedural BBB or PPM undergoing TAVR between 2007 and 2017 were followed for 5 years or until data extraction in September 2017. Data on vital status and hospitalization were obtained through national registries.
Results Within 30 days post-TAVR, new BBB without PPM and new PPM occurred in 247 (30.3%) and 132 (16.2%) patients, respectively, leaving 437 patients (53.6%) without conduction abnormalities. Median follow-up was 2.5 years (interquartile range: 1.0 to 4.9 years). One-year all-cause mortality was increased for new BBB (hazard ratio [HR]: 2.80; 95% confidence interval [CI]: 1.18 to 3.67) but not for new PPM (HR: 1.64; 95% CI: 0.72 to 3.74) compared with patients with no conduction abnormalities. The risk for late all-cause mortality (≥1 year after TAVR) was higher both for patients with new BBB (HR: 1.79; 95% CI: 1.24 to 2.59) and for those with new PPM (HR: 1.58; 95% CI: 1.01 to 2.46) compared with patients with no conduction abnormalities. Patients with new BBB (HR: 1.47; 95% CI: 1.02 to 2.12) and new PPM (HR: 1.66; 95% CI: 1.09 to 2.54) had a higher risk for heart failure hospitalization and reduced left ventricular ejection fraction (p < 0.0001 for both groups) during follow-up.
Conclusions New BBB and new PPM developed frequently after TAVR. New BBB was associated with increased early and late all-cause mortality, whereas new PPM was associated with late all-cause mortality. Furthermore, both new BBB and new PPM increased the risk for heart failure hospitalizations.
Increased operator experience and improved transcatheter heart valve (THV) design are among factors that have reduced the risk for transcatheter aortic valve replacement (TAVR)–related complications (1–3). The association between more than mild paravalvular regurgitation (PVR) and mortality was an early recognized potential Achilles’ heel of TAVR (4). As a consequence, both manufacturers and operators have had a strong focus on reducing the rate of PVR. However, preventive measures against PVR, such as the introduction of sealing skirts, more liberal THV oversizing, and post-dilation, are risk factors for the development of conduction abnormalities (CAs) after TAVR (2,5). This might explain why induction of left bundle branch block (LBBB) and implantation of a permanent pacemaker (PPM) remain frequent complications after TAVR, seen in 10% to 30% of patients depending on the type of THV implanted (1,3,5–8). Still, the clinical impact of TAVR-induced CAs remains controversial, as new-onset LBBB and post-procedural PPM have shown inconsistent effects on mortality and heart failure after TAVR (5,7,9–19).
The aim of the present study was to investigate the incidence of TAVR-induced new-onset bundle branch block (BBB) and new PPM implantation as well as their impact on all-cause mortality and hospitalization for heart failure in an all-comers TAVR population.
The present study was a prospective single-center study. All consecutive patients who underwent TAVR at Rigshospitalet, Copenhagen University Hospital, Denmark, from 2007 to 2017 were included. Patients with pre-procedural BBB or PPM, those with missing pre- or post-procedural electrocardiographic data, and those who emigrated were excluded. The study was approved by the Danish Data Agency and Danish Patient Safety Authority.
Baseline and periprocedural characteristics were collected from electronic patient records. Pre- and post-procedural electrocardiographic data were validated in all patients. QRS interval ≥120 ms was defined as BBB and further classified into LBBB and right BBB (RBBB) (20).
All Danish permanent residents are assigned a unique personal identification number, which is linked to nationwide administrative registries (21). In the Civil Registration and National Patient Registry, vital status and hospitalization information on date of admission and discharge, and codes for diagnosis and procedures are registered. Date and indication of PPM implantation are registered in the Danish Pacemaker Registry.
Patients were classified into 3 groups: no CAs (patients without BBB on the last available 12-lead electrocardiogram or PPM within 30 days after TAVR), new BBB (patients with new-onset LBBB or RBBB on the last available 12-lead electrocardiogram but no new PPM within 30 days after TAVR), and new PPM (patients who underwent PPM implantation within 30 days after TAVR, regardless of the presence or absence of new BBB).
On the basis of the nationwide registries, the primary endpoint was all-cause mortality. Secondary outcomes were first hospitalization and recurrent hospitalizations for heart failure or all-cause admissions. Hospitalization for heart failure was classified if the following International Classification of Diseases-10th Revision codes were registered during admission: I11.0, I13.0, I13.2, I42.0, I42.1, I42.2, I42.9, and I50. Date and indication of PPM implantation were identified through the Danish Pacemaker Registry.
Patients were followed until the date of death or the date of data extraction from the nationwide registries in September 2017, with a maximum of 5 years of follow-up.
Categorical variables are expressed as counts and percentages and were compared using chi-square-tests. Continuous variables are expressed as mean ± SD, compared using analysis of variance and paired Student’s t-tests, or median (interquartile range [IQR]), compared using Kruskal-Wallis tests. Time zero for all survival analyses was set at 30 days post-TAVR. The exposure variable (new BBB or new PPM vs. no CAs) was defined at time zero. The median follow-up time was calculated using the reverse Kaplan-Meier method (22). All-cause mortality was analyzed using the Kaplan-Meier method and the log-rank test. The absolute risk for first heart failure hospitalization (with death without heart failure being a competing risk) was analyzed using the Aalen-Johansen method and Gray’s test.
Cox regression was used to analyze the association of exposure with all-cause mortality rates and rates of first heart failure hospitalization. The models were adjusted for sex, age at date of TAVR as a continuous variable, history of ischemic heart disease, need for dialysis at admission for TAVR, chronic obstructive pulmonary disease, diabetes mellitus, pre-TAVR left ventricular ejection fraction (LVEF), and type of implanted THV. A landmark analysis was performed for all-cause mortality 1 year after TAVR. The Kaplan-Meier estimates were computed among 1-year survivors, and the exposure hazard ratio (HR) was allowed to change after 1 year in Cox regression. The importance of the exposure time interaction (allowing a change of HR after 1 year) was tested using a likelihood ratio test against the proportional hazard model. In further analysis, Cox regression was used to analyze the association of exposure as a time-dependent variable with all-cause mortality during the first 30 days after TAVR. Adjusted mean number of recurrent hospitalizations was estimated as proposed by Ghosh and Lin (23) considering death as a competing risk. For analysis of last available ventricular pacing percentage, a similar cutoff value, of ≤40% and >40%, as in the MOST trial was used (24). The level of statistical significance was set at 5%. All statistical analyses were performed using SAS Base version 9.4 (SAS Institute, Cary, North Carolina).
In total, 1,190 consecutive patients had undergone TAVR from 2007 to September 2017, of whom 348 were excluded because of pre-procedural BBB or PPM, missing electrocardiographic data, or emigration. Of the remaining 842 patients, 26 had <30 days of follow-up after TAVR, leaving 816 patients eligible for the primary analyses. In this final study population, 247 patients (30.3%) had new BBB without new PPM, 132 (16.2%) had new PPM, and 437 (53.6%) had no CAs (Figure 1). The median time to last available electrocardiogram after TAVR was 4 days (IQR: 2 to 7 days) for patients with new BBB or no CAs. PPMs were implanted in 79 patients (59.8%) with new PPM within 1 week after TAVR. New PPM was single-ventricular (VVI/DDD device) in 122 patients (92.4%) and biventricular (cardiac resynchronization therapy device) in 9 patients (6.8%); a total of 3 (2.3%) of these also functioned as implantable cardioverter-defibrillators. More patients with new BBB (n = 19 [7.7%]) received PPMs later than 30 days after TAVR compared with patient with no CAs (n = 15 [3.4%]) (p = 0.013). Median follow-up was 2.5 years (IQR: 1.0 to 4.9 years).
Baseline and procedural characteristics are shown in Table 1. The median age of the total population was 81 years (IQR: 75 to 85 years), and the median Society of Thoracic Surgeons score was 3.2% (IQR: 2.2% to 4.9%). There was no difference in the baseline characteristics between groups, except for the type of implanted THV (p < 0.0001) and baseline LVEF (p = 0.049).
Five years after TAVR, all-cause mortality was 48.4% in patients with new BBB, 46.7% in those with new PPM, and 32.8% in those with no CAs (p = 0.0003) (Figure 2A).
The hazard rate of early (<1 year) and late (≥1 year) all-cause mortality was significantly higher for patients with new BBB compared with those with no CAs, whereas only late all-cause mortality was higher for patients with new PPM compared with no CAs (Figure 2B). The HRs of all-cause mortality during the first 30 days after TAVR were 4.89 (95% confidence interval [CI]: 1.00 to 23.92) for patients with new BBB and 2.21 (95% CI: 0.20 to 24.72) for those with new PPM compared with no CAs.
The hazard rate of first heart failure hospitalization was higher for patients with new BBB and new PPM compared with those with no CAs (Figure 3). In a subanalysis, first heart failure hospitalization was included in the Cox regression model as a time-dependent variable and independently increased the risk for all-cause mortality (HR: 3.32; 95% CI: 2.41 to 4.59; p < 0.0001).
The mean number of recurrent heart failure hospitalizations up to 5 years after TAVR was higher for patients with new BBB and new PPM compared with those with no CAs (Table 2, Online Figure 1). However, when excluding patients without any heart failure admissions, there was no difference in the mean number of recurrent heart failure hospitalizations during follow-up. There was no difference in the mean number of recurrent all-cause hospitalizations for patients with new BBB and new PPM compared with no CAs (Table 2, Online Figure 2).
For patients admitted to the hospital, the median duration of hospital stay was 4.5 days/admission (IQR: 2.0 to 7.0 days/admission) for patients with new BBB, 4.0 days/admission (IQR: 2.0 to 6.0 days/admission) for those with new PPM, and 3.0 days/admission (IQR: 1.7 to 6.0 days/admission) for those with no CAs (p = 0.037).
The 10 patients with new RBBB had a median follow-up duration of 2.9 years; none of these patients died during follow-up. Exclusion of patients with new RBBB from those with new BBB provided similar results for primary analyses of all-cause mortality and heart failure (Online Figure 3).
Data for pacing percentage were available for all patients with new PPM except 1. The last available recorded pacing data for patients with new PPM showed that 48.1% (n = 63) and 51.9% (n = 68) of patients had ventricular pacing percentages ≤40% and >40%, respectively. Both groups had a median backup rate of 60 beats/min (IQR: 60 to 60 beats/min; p = 0.43). In patients with pacing percentages ≤40%, the last available electrocardiogram within 30 days post-TAVR (2 patients had missing data) showed that 41 patients (67.2%) had new-onset LBBB, 9 (14.8%) had new-onset RBBB, and 11 (18.0%) had no BBB when not paced.
Including only patients with single-ventricular pacemakers (VVI/DDD devices), the hazard rate of all-cause mortality was similar for patients with right ventricular pacing (RVP) ≤40% versus >40% (Table 3). The hazard rate of first heart failure hospitalization was higher for patients with RVP >40% compared with those with RVP ≤40% (HR: 2.84; 95% CI: 1.25 to 6.45).
Changes of LVEF
In 320 patients (42.1% of patients at risk), the last available echocardiogram older than 6 months (median time to follow-up 2.0 years [IQR: 1.1 to 4.0 years] with no differences between groups [p = 0.89]) was analyzed for follow-up. LVEF was similar between groups at discharge (p = 0.73) (Figure 4). LVEF was significantly reduced from discharge to follow-up among patients with new BBB and new PPM (p < 0.0001 for both groups) but not for patients with no CAs. LVEF was significantly different between groups at follow-up (p = 0.0001) (Figure 4).
Of 816 patients without BBB or PPM prior to TAVR, 30.3% patients developed new BBB, and 16.2% had need for PPM within 30 days after TAVR. The rates of new BBB (5–7,10–13) and new PPM (1,3,5,7,14–17) were similar to those previously reported.
Compared with patients with no CAs, the risk for early and late all-cause mortality was increased in patients with new BBB, whereas only the risk for late all-cause mortality was increased in patients with new PPM. Furthermore, LVEF was reduced during follow-up, the risk for heart failure hospitalization was increased, and the duration of hospitalization was longer for patients with new BBB or new PPM compared with those with no CAs.
Clinical outcomes of new BBB after TAVR
In a meta-analysis, new LBBB was found to increase the risk for cardiac death and PPM implantation within 1 year after TAVR (7). Further, Urena et al. (6) observed an increased risk for sudden cardiac death for TAVR patients with new LBBB, which increased additionally in an interactive manner during the first 6 months if new LBBB was concomitant with LVEF ≤40%. These results could indicate an increased risk for early progression to heart failure, complete heart block, or ventricular arrhythmias.
LBBB is known to cause interventricular dyssynchrony, shortening of diastole, and abnormal septal motion, reducing LVEF and over time inducing asymmetrical cardiac dilation and hypertrophy, leading to heart failure (25). Several studies have described a reduced immediate and late recovery of LVEF in patients with new LBBB after TAVR (9,11,13) but failed to show an increased risk for heart failure hospitalizations (9,12). In the present study, heart failure hospitalization was found to be a strong predictor of all-cause mortality, and the increased risk for heart failure hospitalization for patients with new BBB might explain the increased risk for late all-cause mortality after TAVR compared with patients with no CAs.
Several of the studies that failed to show an association between new LBBB and all-cause mortality defined new LBBB as LBBB present at discharge, included patients with new PPM >48 h post-TAVR or after discharge in the comparator group, and had higher baseline risk score compared with the present study (9,11–13). The inclusion of a higher proportion of patients with transient LBBB and a higher weight of comorbidities may attenuate the difference in clinical outcome in these studies.
Unlike patients with pre-procedural RBBB, limited follow-up data exist for patients with new RBBB after TAVR. Similar to what has previously been observed (11), 1.2% of patients developed new RBBB after TAVR in the present study. Although none of these patients died during follow-up, further data are needed.
Clinical outcomes of new PPM after TAVR
In the present study, patients with new PPM only had increased risk for late all-cause mortality, potentially being protected from aforementioned risk for early sudden cardiac death seen in patients with new LBBB (6). In the meta-analysis by Regueiro et al. (7), new PPM after TAVR failed to show an association with all-cause mortality but had a tendency to protect from cardiac death up to 1 year after TAVR. However, the analyzed studies included patients with both pre-procedural and new BBB (7,15,17,18). It is plausible that the observed protective effect of new PPM was because a significant proportion of the patients without PPM after TAVR had new LBBB, which was associated with 1-year cardiac death (7).
Previous studies with more than 1 year of follow-up failed to show an association between new PPM after TAVR and all-cause mortality. However, these studies included patients with known or new LBBB in the comparator group (14,15). RVP is known to cause similar interventricular dyssynchrony, resulting in an analogous pathway to heart failure as seen in patients with LBBB (25). Given that LBBB and RVP increase the risk for heart failure, the inclusion of patients with LBBB in the comparator group might be why previous studies did not identify an association between new PPM after TAVR and longer-term mortality (14,15).
Outcome of prolonged QRS duration
As the indication for TAVR advances to younger patients with low surgical risk and longer life expectancy, the long-term safety of TAVR becomes increasingly important. In the present study, PPM implantation seemed beneficial in case of new CAs, maybe protecting from early sudden cardiac death. However, the increased risk for heart failure hospitalization and low LVEF at follow-up for both patients with new BBB and new PPM could be a cofactor in the increased risk for late mortality observed in these patients compared with those with no CAs. Thus, neither new BBB nor new PPM appears benign in the long term, indicating that prevention is the best long-term treatment of TAVR-induced CAs and that these patients may benefit from closer follow-up. Balloon-expandable and some self-expanding THVs tend to have a lower risk for CAs (5), but, the risk for other complications such as PVR also needs to be considered when selecting the optimal THV. Reducing the risk for CAs with a minimal trade-off for an increased risk for PVR, the membranous septum length might function as patient-specific limit for THV implantation depth (26).
Similar to the results of the MOST trial (24), patients in the present study with new PPM who were paced >40% of the time had higher risk for heart failure hospitalizations compared with those paced ≤40%. Periodic reassessment of the intrinsic QRS configuration and heart rate for patients with PPM at follow-up could reveal some patients with unnecessarily high RVP percentages who might benefit from a lower paced backup rate. Cardiac resynchronization therapy has been shown to be beneficial in case reports of patients with low LVEFs and persistent LBBB after TAVR (5). However, there are limited data on the systematic implementation of cardiac resynchronization therapy devices for TAVR patients with reduced LVEFs and even more so for preventing deterioration of a normal LVEF, which might be increasingly important for younger low-risk patients with TAVR-induced CAs.
The diagnosis of heart failure was based on data from national registries, which contain information on the diagnosis given by the treating physician at discharge. It was not possible to validate diagnoses; however, the national registries provide unselected information, which is mandatory to complete, on all patients every time they are admitted to the hospital. A previous study validating the Danish National Patient Registry found a positive predictive value of 75% for heart failure hospitalizations (21).
The median time to last available electrocardiogram was 4 days, meaning that some patients might have been misclassified because of only transient BBB. However, analysis with time zero set at 3 months, giving a median time to last available electrocardiogram of 7 days (IQR: 3.0 to 35.0 days), provided similar results for all-cause mortality and heart failure hospitalization (Online Figure 4).
Even though baseline differences were minimal and most analyses were multivariate, the observational nature of the study makes it possible that unmeasured confounders might explain the observed results, and external validity might be limited because of the single-center design of the study.
Patients with new BBB had an increased risk for early all-cause mortality compared with those with no CAs. The risk for heart failure hospitalization and late all-cause mortality was increased, and the duration of hospitalization was longer for patients with new BBB or new PPM compared with those with no CAs. Although future larger scale, multicenter studies should confirm these results, this study indicates that TAVR-induced new BBB and new PPM are not just benign post-TAVR complications.
WHAT IS KNOWN? New LBBB and new PPM are frequent complications after TAVR. However, previous studies have reported inconsistent results on their clinical impact.
WHAT IS NEW? Both new BBB and new PPM after TAVR are associated with increased all-cause mortality, although new PPM may protect against early death. Similarly, TAVR-induced CAs increase the rate of heart failure hospitalization as well as the duration of hospitalization in general.
WHAT IS NEXT? Future studies should confirm these results, separating patients with new-LBBB, those with new PPM, and those without CAs after TAVR. Furthermore, procedural planning and selection of THV may reduce the risk for the development of CAs without increasing the risk for PVR, which might be of significant value for long-term outcome after TAVR.
Dr. Jørgensen has received a research grant from Edwards Lifesciences. Dr. Sondergaard has received consulting fees and institutional research grants from Abbott, Boston Scientific, Edwards Lifesciences, Medtronic, and Symetis. Dr. Svendsen is on the advisory board of Medtronic; has received speaking fees from Medtronic and Biotronik; and has received an institutional research grant from Medtronic, Biotronik, and Gilead. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- bundle branch block
- conduction abnormality
- confidence interval
- hazard ratio
- interquartile range
- left bundle branch block
- left ventricular ejection fraction
- permanent pacemaker
- paravalvular regurgitation
- right bundle branch block
- right ventricular pacing
- transcatheter aortic valve replacement
- transcatheter heart valve
- Received August 13, 2018.
- Revision received October 25, 2018.
- Accepted October 30, 2018.
- 2019 American College of Cardiology Foundation
- Auffret V.,
- Lefevre T.,
- Van Belle E.,
- et al.
- Cahill T.J.,
- Chen M.,
- Hayashida K.,
- et al.
- Grover F.L.,
- Vemulapalli S.,
- Carroll J.D.,
- et al.
- Généreux P.,
- Head S.J.,
- Hahn R.,
- et al.
- Auffret V.,
- Puri R.,
- Urena M.,
- et al.
- Urena M.,
- Webb J.G.,
- Eltchaninoff H.,
- et al.
- Regueiro A.,
- Abdul-Jawad Altisent O.,
- Del Trigo M.,
- et al.
- Siontis G.C.M.,
- Jüni P.,
- Pilgrim T.,
- et al.
- Urena M.,
- Webb J.G.,
- Cheema A.,
- et al.
- Houthuizen P.,
- Van Garsse L.A.F.M.,
- Poels T.T.,
- et al.
- Testa L.,
- Latib A.,
- De Marco F.,
- et al.
- Carrabba N.,
- Valenti R.,
- Migliorini A.,
- et al.
- Chamandi C.,
- Barbanti M.,
- Munoz-Garcia A.,
- et al.
- Urena M.,
- Webb J.G.,
- Tamburino C.,
- et al.
- Fadahunsi O.O.,
- Olowoyeye A.,
- Ukaigwe A.,
- et al.
- Nazif T.M.,
- Dizon J.M.,
- Hahn R.T.,
- et al.
- Mouillet G.,
- Lellouche N.,
- Yamamoto M.,
- et al.
- Buellesfeld L.,
- Stortecky S.,
- Heg D.,
- et al.
- Surawicz B.,
- Childers R.,
- Deal B.J.,
- Gettes L.S.
- Sundbøll J.,
- Adelborg K.,
- Munch T.,
- et al.
- Sweeney M.O.
- Hamdan A.,
- Guetta V.,
- Klempfner R.,
- et al.