Preprocedural and procedural variables that predict new-onset conduction disturbances after transcatheter aortic valve replacement

Transcatheter aortic valve replacement (TAVR) is an emerging therapeutic procedure in patients with symptomatic severe aortic stenosis who have intermediate to high surgical risk or who are inoperable [1‐5]. Despite the development of the transcatheter heart valve (THV) to decrease periprocedural complications, new-onset conduction disturbances post-TAVR (hereafter collectively referred to as CD/CDs) remain relatively common [6]. The mechanism of CDs may be direct mechanical insult by the THV to conduction tissue, which results in ischemia, edema, or hemorrhage [6]. Based on this mechanism hypothesis, several studies have investigated the predicted risk for permanent pacemaker implantation (PPI) in patients with CDs, but some of the reported predictors were inconsistent among studies [7, 8]. Because of these variations among treatment-related factors, including discrepancies in the indication and timing of PPI, the care decision pathway has not yet been conclusively established in this clinical setting. Moreover, the data are still scarce relative to the clinical outcomes of patients with CDs compared between those treated with and without PPI [9]. It is possible that there may be a single standardized decision pathway that is suitable for use in CD patients.

Accordingly, the aim of this study was to investigate clinical, electrocardiographic, preprocedural imaging, and procedural variables to identify independent factors that predict new-onset conduction disturbances post-TAVR.

During December 2009 to March 2021, 132 patients underwent TAVR. After exclusion of patients with implantation of cardiac device at baseline (n = 6, 4.5% of all patients), or intraoperative death (n = 2, 1.5% of all patients), the remaining 124 patients were enrolled. Of those, 35 (28.2%) patients developed CDs, including 19 (54.3%) new-onset LBBB, 13 (37.1%) new-onset significant CD, and 3 (8.6%) new-onset high-grade AV block. The overall rate of permanent pacemaker implantation was 9.7% (12 of 124 cases). Preprocedural and procedural clinical and ECG characteristics are shown in Tables 2, 3 and 4.

The present study investigated preprocedural and procedural variables to predict a new-onset conduction disturbance post-TAVR (CD/CDs). The main findings of this study are, as follows: (1) The prevalence of CD development was 28.2%, and PPI was required in 9.7%.; (2) The preprocedural characteristics prior IVCD, MAC, lower LV mass index, and MS length \(\le\) 6.43 mm were more likely to be present in patients with a CD; (3) Procedural characteristics of the use of a self-expandable TAVR, smaller LV cavity during the procedure, and ID > 6 mm were also more commonly present in patients with a CD; (4) Preprocedural MS length and postprocedural difference between MS length and ID (∆MSID) was found to be the strongest predictors of a CD; (5) An ∆MSID cut-off value of < 0 mm was found to most strongly predict a CD. Importantly and different from any previous study, we expanded the criteria for defining a CD to cover all spectrum of TAVR-related conduction injury to lower the threshold of this only modifiable risk factor.

This study found a CD rate of 28.2%, which corresponds with the previously reported incidence of new LBBB after TAVR (range 20% to 34%) [8, 19]. One-third of our patients with CD were treated by PPI. We broadened our study CD endpoint to include new-onset complete LBBB and/or RBBB, significant IVCD (QRS duration ≥ 120 ms), and new-onset HAVB or CHB for cover the entire spectrum of TAVR-related conduction injury.

Our study categorized variables into preprocedural and procedural predictors, and we included many parameters that were reported to be associated of conduction disturbances post-TAVR. This study included clinical, electrocardiographic, echocardiographic, MSCT, and procedural data for analysis. Transapical access was reported to be less likely associated with new-onset CD post-TAVR due to the use of a balloon expanding device [20]. Parameters, including intraventricular conduction delay, mitral annular calcification, MS length ≤ 6.43 mm, lower LV mass index, small LV cavity, use of self-expanding device, ID ≥ 6 mm, were analyzed for their ability to significantly/independently predict CD. We also included other variables that was previously report [8] in predicting CDs in the analysis. The results of our multivariate analysis revealed MS length ≤ 6.43 and ∆MSID < 0 as independent predictors of CDs. The lower border part of the MS is the typical location of the atrioventricular bundle [18]. Shorter MS also correlates with high position or less ventricular position of the anatomical atrioventricular bundle. MS length is easily measured with good reproducibility in the coronal plane [21]. Mauri, et al. first described the usefulness of angiographic implantation depth, device landing zone calcium volume, and baseline RBBB as independent predictors of the need for PPI post-TAVR [22]. Maeno, et al. identified non-coronary cusp device landing zone calcium volume, RBBB, and ∆MSID as synergistic PPI predictors [23]. The usefulness of calcium in device landing zone and implantation depth in predicted PPI in balloon expanding device have been demonstrated [24]. Tretter, et al. analyzed gross anatomical variation in the aortic root, such as the distance from the virtual basal ring to the inferior margin of the MS, relative to correlation with PPI or LBBB. They found TAVR valve type and ID to be the primary procedural factors associated with conduction damage regardless of gross anatomical variation [25]. Concerning TAVR valve type, the Sapiens 3 and Accurate Neo valves were reported to be associated with the lowest PPI rate, followed by the Evolut and Portico valves [26]. Jilaihawi, et al. proposed an anatomically-guided minimized depth during the MS approach for implantation of a self-expanding TAVR [15]. Chen, et al. reported an ∆MSID of 3.2 mm to be the optimal cutoff point for predicting a CD when using a self -expanding device [21]. For balloon expanding TAVR, mean ∆MSIDs of − 2.5 ± 2.4 mm [23], 0.5 ± 4 mm [24], and − 1.7 ± 1.5 mm [27] were reported in patients with PPMI or LBBB. Miki, et al. reported that an ∆MSID cut-off value of -0.7 mm had an AuROC of 0.85, sensitivity of 0.78, and specificity of 0.82 (p < 0.001). That group defined a CD as an atrioventricular block that required PPI or new-onset complete LBBB. Because the ∆MSID was found to be the only modifiable predictor, we expanded our definition of CD to include PPI, new onset-LBBB, onset-RBBB, significant IVCD (QRS duration ≥ 120 ms), and new-onset HAVB or CHB for comprehensively cover the spectrum of TAVR-related conduction injury. In our study, an ∆MSID cut-off value of < 0 mm was found to be the optimal cutoff point for predicting a CD (AuROC: 0.896).

Factors, such as baseline RBBB, amount of calcium at the device landing zone, and MS length, are not modifiable. We, therefore, lowered the threshold of the only modifiable predictor (∆MSID) to cover the entire spectrum of TAVR-related conduction injury, which has the effect of lowering the risk of PPI. Using an ∆MSID cut-off value of < 0 mm (i.e., attempt to deploy the THV at an ID less than the MS length) is the best way to minimize the of CDs.

Many techniques have introduced to minimize implantation depth. Cusp-Overlap view strategy recommended for self-expandable valve implantation [28, 29]. For balloon-expandable SAPIEN-3 (S3) valve, the systematic approach to high deployment technique of SAPIEN-3 valve reduced 30-day permanent pacemaker rates from 13.1% to 5.5% [30]. However, this systemic approach to high implantation may constraint future coronary access in certain anatomical subtypes such as low coronary height, narrow coronary sinus, low and narrow sinotubular junction. The risk of coronary obstruction, future coronary access and new-onset CDs post-TAVR can be balanced by knowing anatomy and its predictors.

This study confirmed the usefulness of immediate postprocedural 12-lead ECG as a screening tool for significant ECG change, including new-onset LBBB and/or increase in PR or QRS duration ≥ 20 ms and/or development of periprocedural transient or persistent heart block, as an indicator of the occurrence of new-onset of conduction disturbances [9, 31]. We also reported a late conduction disturbance rate at 1.6%, which is similar previous reports (range 2% to 7%) [6, 32]. More specifically, we had only 1 patient with no significant peri- or immediate postprocedural ECG change that developed late high-degree conduction disturbances with AF with bradycardia during follow-up.

The rate of pacemaker dependence reported in previous studies ranged from one- to two-thirds of patients during follow-up [33‐35]. New-onset LBBB post-TAVR resolved in 45% (range 35–56%) in 30 days [6]. Our study found a lower rate of pacemaker dependency (16%) with a similar rate of conduction disturbance resolution (55%). This difference between studies may be explained by a higher unnecessary PPI rate. The most common indication for PPI according to the judgment of our electrophysiologist was new-onset LBBB. Further study is needed to identify the appropriate indication(s) for PPI in this clinical setting.

The results of this study revealed preprocedural MSCT-derived MS length and procedural ID to be independent predictors of new-onset conduction disturbance post-TAVR. An ∆MSID cut-off value of < 0 mm was found to optimally predict a CD. Importantly and in contrast to previous studies that investigated and reported a limited criteria for defining a CD, we investigated the full spectrum of possible causes of CD, including PPI, new onset-LBBB, onset-RBBB, significant IVCD (QRS duration ≥ 120 ms), and new-onset HAVB or CHB, because the ∆MSID was found to be the only modifiable predictor. This insight encourages routine deployment of the THV at an ID less than the MS length.