Predictors of Response to Left Bundle Branch Area Pacing from CMR with Multiple Imaging Sets: A Multicenter Study (RF_TH_146)

Left bundle branch area pacing (LBBAP) is an emerging pacing therapy for heart failure and is increasingly used as an alternative to cardiac resynchronization therapy; however, the mechanical effects of LBBAP on the left ventricle (LV) and right ventricle (RV), and the dependence of these effects on procedural results are poorly understood. As CMR volumetric imaging has greater reproducibility compared to other imaging modalities, CMR in the chronic phase after pacemaker implantation promises to result in important mechanistic and procedural insights for LBBAP. Furthermore, assessment of changes in ventricular function with pacing on versus pacing off is now feasible based on our published protocol (see References) for scanning multiple imaging sets corresponding to different pacing settings during a single CMR examination.

Methods:

In a multicenter study with the University of Virginia and the Virginia Commonwealth University, patients with and without heart failure with implanted pacemakers having LBBAP underwent CMR scans with balanced steady-state free precession (bSSFP) cine imaging and up to three imaging sets. Late gadolinium enhancement and strain imaging were also performed in many patients. Patients with intact atrioventricular conduction had an imaging set with atrial-only pacing and no LBBAP to demonstrate the acute effect of turning LBBAP on and off. The 12-lead ECGs corresponding to the pacing settings scanned were analyzed with respect to the QRS morphology in leads V1 and V6, QRS duration with and without latent conduction, and the R-wave peak time in lead V6.

Results:

Among 39 patients (age 73.5 ± 8.3 years; 33.3% female), 25 (64.1%) had intact AV conduction and also underwent an imaging set with LBBAP off in addition to LBBAP on. Pacing in two of the 39 patients overall was determined capture the His bundle based on standard criteria. In all 39 patients, the QRS duration with LBBAP was moderately correlated with the LVEF with pacing (r=-0.39, p=0.019, Figure 1A) but not significantly correlated with the RVEF with pacing (r=-0.24, p=0.18, Figure 1B). Of the 25 patients with intact AV conduction, 12 (48%) had improved LVEF with LBBAP, 6 (24%) had no change in the LVEF with LBBAP, and 7 (28%) had decreased LVEF with LBBAP. The change in RV and LV end-systolic volume indices (RVESVI and LVESVI, respectively) with LBBAP on versus off were highly correlated (r=0.62, p=0.0023, Figure 1C). The best predictor of improvements in the LVESVI and RVESVI with LBBAP was the presence of a completely positive QRS (R wave morphology) in lead V6, indicating that the LBBAP lead had been implanted in the more proximal (versus distal) conduction system (p=0.0098 for LVESVI and p=0.0012 for RVESVI, Figure 2). Conversely, the change in these CMR volumetric parameters with LBBAP on versus off were similar in patients with and without a late R wave in lead V1, an ECG finding that indicates capture of the left posterior aspect of the interventricular septum (p=0.49 for LVESVI and p=0.14 for RVESVI).

Conclusion:

This first multicenter study of CMR with multiple imaging sets in patients with LBBAP demonstrated important mechanistic findings that could help improve implementation of this pacing approach for heart failure. In particular, the paced QRS duration was only moderately correlated with ventricular function during pacing, and a ventricular pacing location in the more proximal conduction system (R wave morphology in V6) predicted greater improvements in LV and RV function with pacing.

Figure 1 – Left and Right Ventricular Function After LBBAP Relative to the Paced QRS Duration. (A) The LVESVI with pacing is moderately correlated with the paced QRS duration (r=-0.39, p=0.019). (B) The RVESVI with pacing was not significantly correlated with the paced QRS duration (r=-0.24, p=0.18). (C) The changes in the LVESVI and RVESVI with LBBAP are highly correlated (r=0.62, p=0.0023).

Figure 2 – Change in LVESVI and RVESVI Relative to QRS Morphology in Lead V6. (A) Patients with a completely positive QRS (R wave morphology) in lead V6 had more favorable reductions in the LVESVI with LBBAP compared to patients having QRS complexes in lead V6 with a negative component (rS, Rs, RS, QS) (p=0.0098). (B) There was also a more favorable reduction in the RVESVI in patients with a completely positive QRS complex (R wave morphology) in lead V6 (p=0.0012).

Figure 3 – Change in LVESVI and RVESVI Relative to QRS Morphology in Lead V1. (A) Patients with a late R wave in lead V1 (indicating capture of tissue on the left posterior aspect of the interventricular septum) had similar reductions in the LVESVI with LBBAP compared to patients who did not have this finding (p=0.49). (B) The change in the RVESVI with LBBAP was also similar in these two groups of patients (p=0.14).