Introduction
Transcatheter aortic valve replacement (TAVR) has been established as standard therapy in patients with moderate and high surgical risk with symptomatic aortic stenosis (AS) and is increasingly preferred over surgical aortic valve replacement (SAVR).1,2 The balloon-expandable transcatheter heart valves (THVs), such as Edwards Lifesciences’ SAPIEN 3 and 3 Ultra, have demonstrated significant advancements in minimising procedural complications, particularly in reducing moderate or severe paravalvular regurgitation and major vascular complications, and have exhibited good clinical outcomes at 30 days and 12 months after implantation.3,4 However, new permanent pacemaker implantation (PPI) remains a potential post-procedural complication after TAVR, which can lead to decreased survival rates, increased care costs, extended hospital stays, mortality and rehospitalisation.5–8
The requirement of new PPI following TAVR using the Edwards SAPIEN 3 THV has been linked to various factors, including pre-existing conduction disturbance, aortic valve calcification, complete right bundle branch block (RBBB), prolonged QRS duration, a short membranous septum and greater implantation depth or valve oversizing.8,9 Also, the latest study using the CONDUCT trial registry reported that valve oversizing was a strong procedure-related risk factor for PPI after TAVR, and there is a limited number of studies that have reported on the clinical outcomes of PPI after TAVR that have produced conflicting results.10–14 Therefore, we undertook a 1-year comprehensive study to analyse the clinical outcomes of PPI versus no PPI after TAVR. Propensity score matching was employed to enhance the validity of our observational study by minimising confounding factors and improving the comparability of the groups.
Methods
The CONDUCT trial was a prospective, multicentre, observational registry that collected data on patients who underwent balloon-expandable TAVR using the Edwards SAPIEN 3 THV at four different European centres: the Heart and Diabetes Centre North Rhine-Westphalia, Germany; Amsterdam University Medical Center, the Netherlands; University Hospital Tübingen, Germany and Linköping University Hospital, Sweden. These sites were selected based on their previous experience with TAVR and recommendations from the trial’s steering committee. The registry was conducted following the principles of the Declaration of Helsinki and complied with local laws and regulations. Approval from the ethics committee at each site was obtained, and all participants gave written informed consent before enrolment.
Objectives
The primary objective was to compare the incidence and risk of clinical outcomes such as all-cause mortality, cardiovascular death, stroke or transient ischaemic attack (TIA) and endocarditis in patients receiving PPI compared with those who did not have PPI after TAVR at 1-year follow-up.
The secondary objective was to compare the major adverse cardiac events (MACE) defined as a composite of all-cause mortality, rehospitalisation due to congestive heart failure (CHF) and stroke/TIA in PPI patients who underwent ≤40% right ventricular (RV) pacing versus patients with >40% RV pacing during 1-year follow-up. The threshold for RV pacing of 40% was selected on the basis of previously observed correlations between pacemaker pacing percentages exceeding this threshold and adverse outcomes.15,16 Propensity score matching was applied to minimise the effects of confounders and improve comparability between groups.
Patient Population
A high-risk cohort of patients was identified based on the results of a retrospective analysis of patients undergoing TAVR at the selected facilities.17 Patients were prospectively enrolled in our registry if they underwent a successful transfemoral TAVR using a balloon-expandable heart valve (Edwards SAPIEN 3 or 3 Ultra) and had at least one of the previously recognised risk factors for PPI, such as pre-existing conduction disturbances including RBBB, left anterior hemiblock, atrioventricular block, prolonged QRS duration and bradycardia, a heavily calcified left ventricular outflow tract (LVOT), or a short membranous septum.
The registry excluded patients with a previous pacemaker, a pre-existing indication for PPI before undergoing TAVR and those receiving valve-in-valve implantation. The heart team at each respective site made the decision to perform TAVR and followed the established local protocol.
Documentation and Endpoints
Baseline information regarding patient demographics, medical history, symptoms, echocardiography and surgical risk scores using EuroSCORE II and the Society of Thoracic Surgeons (STS) risk score was documented. 1-year follow-up data of patients undergoing PPI and no PPI after TAVR were recorded to assess the clinical outcomes such as all-cause and cardiovascular mortality, stroke or TIA and endocarditis. RV stimulation was recorded on routine pacemaker check-ups in patients with PPI.
Statistical Analysis
Descriptive statistics were used to analyse the data, with categorical variables presented as absolute values and percentages, while continuous variables were presented as means and standard deviations or medians with interquartile range (IQR). The percentages were calculated based on the number of patients with valid data per parameter, i.e. excluding patients with missing information.
Propensity scores (PS) were calculated to assess PPI-specific effects using selected covariates: sex; hypertension; age; coronary disease; renal insufficiency; CHF; cerebrovascular accident; diabetes; coronary artery disease; previous coronary artery bypass graft; left ventricular ejection fraction (LVEF); and aortic valve peak pressure gradient. The 1:4 ratio PS matching was performed using nearest neighbour matching with a calliper width equal to 0.2 times the standard deviation of the PS logit. Post-matching, standardised mean differences were analysed for all covariates included in the PS calculation. The mean differences for all covariates (except CHF) post-matching were within a desirable threshold (± 0.1), indicating adequate balance (Supplementary Figure 1).
Unadjusted and multivariable-adjusted analyses were performed using Cox proportional hazards regression. In the multivariable analysis, relevant variables in Table 1 with a p<0.1 were included in the model. Relative risks were expressed in the HR and 95% CI. Survival curves were constructed with the Kaplan-Meier analysis. All statistical analyses were performed using R Core Team (https://www.R-project.org/). A significance level of p<0.05 was considered statistically significant.
Results
Of the 295 patients enrolled in the registry, 39 patients who underwent PPI and 256 patients who did not undergo PPI within 30 days post-TAVR were included in the entire cohort. At baseline – a propensity score-matched (PSM) cohort was established using 1:4 nearest-neighbour matching with a predefined calliper, resulting in a total of 160 matched pairs, with 36 patients receiving PPI and 124 patients not receiving PPI after TAVR (Figure 1).
Patient Characteristics and Echocardiography
In both entire and PSM cohorts, patients had a median age of 80 years, were more likely to be men (80.6% and 84.7% patients with and without PPI in the PSM cohort, respectively), and had a median BMI of about 27 kg/m2 (Table 1). At baseline in the PSM cohort, EuroSCORE II (median 2.3% versus 2.1%) and STS risk scores (median 2.1% versus 2.2%) were comparable in patients who did and did not undergo PPI after TAVR. In the entire cohort, type 2 diabetes was more prevalent in patients who underwent PPI (41.0% versus 26.2%; p=0.055); these differences were less pronounced after PSM (38.9% versus 27.4%; p=0.186).
Echocardiography findings showed patients with PPI had a lower LVEF (53% versus 55%; p=0.034). However, this difference vanished after PSM. Patients with PPI had higher systolic pulmonary arterial pressure than patients without PPI (49 mmHg versus 39 mmHg; p=0.013), which remained significantly higher after PSM (49 mmHg versus 38 mmHg; p=0.035). Comparison of other echocardiographic parameters, such as aortic, mitral and tricuspid regurgitation, aortic valve area (planimetry and indexed), and aortic valve pressure gradient, indicated no significant difference between patients who did or did not undergo PPI in both the entire and PSM cohorts. Further, there was no significant difference in the history of coronary artery bypass graft or valve surgery. The characteristics of patients who received pacemakers (n=28) in the PSM cohort and experienced RV pacing ≤40% versus >40% showed no significant differences in age, surgical risk and comorbidity burden, and the only significant difference was the higher prevalence of type 2 diabetes in patients who experienced >40% RV pacing (52.2% versus 0%; p=0.053; Supplementary Table 1).
One-Year Clinical Outcomes
At 1-year follow-up after TAVR, the incidence of MACE after PSM was higher in patients with PPI than in patients without PPI, however, this difference was not statistically significant (22.2% versus 13.7%; p=0.216; Table 2 and Supplementary Figure 2) (HR 1.63; 95% CI [0.72–3.71]; Table 3). However, rehospitalisation due to worsening CHF was significantly higher in patients receiving PPI (11.1% versus 2.4%; p=0.046) (HR 5.05; 95% CI [1.09–23.4]). The Kaplan-Meier analysis on 1-year freedom from composite MACE revealed no difference in patients with or without PPI (p=0.248; Figure 2). In the full cohort, unadjusted outcomes were worse compared to the PSM cohort; for example, the incidence of MACE was significantly higher in patients with PPI than in patients without PPI (23.1% versus 11.3%; p=0.041; Table 2) (HR 2.26; 95% CI [1.11–4.64]; Supplementary Table 2). This was mainly driven by rehospitalisation due to worsening CHF, which was significantly higher in patients receiving PPI (12.8% versus 2.3%; p=0.008) (HR 5.85; 95% CI [1.77–19.4]; Supplementary Table 2).
We analysed long-term RV stimulation in 28 patients in the PSM cohort. The mean RV stimulation was 78.6%. High RV stimulation (>40%) was observed in 23 patients (82%), while five patients (18%) had low RV stimulation. One patient recovered from AV block and no longer required RV pacing. There was no significant difference in the occurrence of relevant clinical events between the low and high RV pacing groups (Figure 3 and Supplementary Table 3). However, the analysis may be underpowered due to the low number of events, with only four events occurring in the low RV stimulation group.
Discussion
Our study examined the 1-year clinical outcomes in matched groups of patients who did or did not undergo PPI within 30 days after TAVR. The results revealed comparable clinical outcomes in terms of all-cause mortality, cardiovascular mortality, stroke or TIA and endocarditis between the two groups. However, there was a higher risk of rehospitalisation due to CHF in patients who underwent PPI. The majority of patients with PPI had high RV stimulation, with only one patient recovering from AV block and no longer requiring RV pacing. The analysis comparing low and high RV stimulation was underpowered and showed no statistically significant difference between the groups.
Due to technical improvements and increased experience in implantation techniques, the PPI rate after TAVR has decreased in recent years, especially with the use of balloon-expandable valves with a sealing skirt. In our registry, which included high-risk patients for high-grade conduction abnormalities, only 14% required PPI after balloon-expandable TAVR.10 However, PPI remains a relevant issue after TAVR, particularly when younger, low-risk patients are treated, as they could be exposed to potential complications over many years of cardiac pacing. Lee et al. provide a comprehensive review of device- and patient-specific clinical predictors for PPI after TAVR, including pre-existing RBBB and prosthesis to LVOT diameter, especially for SAPIEN valves.18 The latter could be used during preoperative assessment and might help to mitigate the probability of provoking heart block. Additionally, the review points to the detrimental effects of long-term RV pacing on LV function. However, the present study’s cohort did not demonstrate any indication of affected LV function by RV pacing at 1-year following TAVR. PPI after TAVR may affect the clinical outcome of patients through several mechanisms. Pacemaker or lead-related complications may occur during implantation, such as bleeding, pneumothorax, wall perforation, and during long-term follow-up, such as lead dysfunction, infection, interference with tricuspid valve resulting in tricuspid regurgitation, but are relatively rare in current practice.19 In our study, we did not observe any peri-interventional complications or increased incidence of endocarditis (one occurrence of pacemaker-related endocarditis) in PPI patients. More importantly, chronic RV pacing may increase the risk of the development or worsening of heart failure symptoms due to its deleterious effect on synchronous ventricular contraction.15 In our study, we found a significant increase in CHF rehospitalisation in PPI patients. However, we did not observe differences in other clinical endpoints such as mortality or stroke. Most patients in our cohort were on high long-term RV stimulation. Physiological pacing may reduce the effects associated with conventional RV pacing.20
Several studies found no difference in the clinical outcome of patients with or without PPI after TAVR. Rück et al.’s nationwide, population-based cohort study of 3,420 TAVR patients reported similar survival rates up to 5-years follow-up and comparable incidences of cardiovascular death, heart failure and endocarditis (median follow-up of 1.8 years) in patients undergoing PPI or no-PPI after TAVR.11 Hochstadt et al. reported that post-procedure PPI (within 30 days) is not associated with an increase in all-cause mortality (up to 6-years follow-up) in TAVR patients.21 Similarly, a meta-analysis by Regueiro et al. also indicated PPI within 30 days after TAVR did not increase the risk of all-cause mortality (RR 1.03; 95% CI [0.90–1.18]; p<0.64) and cardiovascular death (RR 0.78; 95% CI [0.60–1.03]; p=0.06) and showed a protective benefit from cardiac death.22
However, in contrast to these findings, a systematic review by Faroux et al. indicated an increased risk of all-cause mortality (RR 1.17; 95% CI [1.11–1.25]; p<0.001) among patients undergoing PPI.12 Furthermore, they also reported an increased risk of HF hospitalisation (RR 1.18; 95% CI [1.03–1.36]; p=0.02) due to pacemaker implantation.12 This finding aligns with our study results showing a significant increase in CHF rehospitalisation in patients with PPI compared to those without PPI (12.8% versus 2.3%; p=0.008; HR 5.85; 95% CI [1.77–19.4]). However, in our study, this increased risk of valve-related or HF rehospitalisation did not affect all-cause or cardiovascular mortality at 1 year between the groups. Similar findings were reported by Jorgensen et al., who also observed increased HF rehospitalisation in patients undergoing PPI without a significant increase in 1-year all-cause mortality.13
In conclusion, clinicians should be aware that worsening heart failure may be relevant after PPI in TAVR patients. Further improvements in valve prosthesis development and implantation techniques are needed to prevent the occurrence of conduction disturbances. Patients with PPI after TAVR should be closely followed up and evaluated for signs and symptoms of heart failure. Intensification of heart failure therapy or even an upgrade to resynchronisation therapy in cases of a high proportion of RV pacing may be appropriate.
Limitations
This is an observational study on patients undergoing TAVR in four experienced European centres. Therefore, the inherent limitations to observational registries, such as minor incompleteness or inaccuracies in patient data, are unavoidable, although measures, such as a carefully designed care report form, data monitoring and verification were performed to minimise such instances. However, the study has a high level of completeness of patient data for up to 1 year, making it a valuable source of evidence for the evaluation of TAVR patients with or without PPI in a real-world setting.
Conclusion
Our study revealed similar 1-year clinical outcomes, such as all-cause mortality, cardiovascular death, stroke or TIA, and endocarditis, in matched groups of patients who did or did not undergo PPI after TAVR. However, there was a higher incidence of rehospitalisation due to valve-related problems or heart failure in patients undergoing PPI after TAVR.
Clinical Perspective
- Permanent pacemaker implantation (PPI) is a complication of transcatheter aortic valve replacement (TAVR), and there is limited evidence of its impact on clinical outcomes.
- This study showed that 1-year clinical outcomes, such as all-cause mortality, cardiovascular death, stroke and endocarditis, did not differ between patients who did or did not undergo PPI after TAVR, indicating a lower short-term impact of pacemaker implantation.