His Resynchronization Versus Biventricular Pacing in Patients With Heart Failure and Left Bundle Branch Block

Background His bundle pacing is a new method for delivering cardiac resynchronization therapy (CRT). Objectives The authors performed a head-to-head, high-precision, acute crossover comparison between His bundle pacing and conventional biventricular CRT, measuring effects on ventricular activation and acute hemodynamic function. Methods Patients with heart failure and left bundle branch block referred for conventional biventricular CRT were recruited. Using noninvasive epicardial electrocardiographic imaging, the authors identified patients in whom His bundle pacing shortened left ventricular activation time. In these patients, the authors compared the hemodynamic effects of His bundle pacing against biventricular pacing using a high-multiple repeated alternation protocol to minimize the effect of noise, as well as comparing effects on ventricular activation. Results In 18 of 23 patients, left ventricular activation time was significantly shortened by His bundle pacing. Seventeen patients had a complete electromechanical dataset. In them, His bundle pacing was more effective at delivering ventricular resynchronization than biventricular pacing: greater reduction in QRS duration (−18.6 ms; 95% confidence interval [CI]: −31.6 to −5.7 ms; p = 0.007), left ventricular activation time (−26 ms; 95% CI: −41 to −21 ms; p = 0.002), and left ventricular dyssynchrony index (−11.2 ms; 95% CI: −16.8 to −5.6 ms; p < 0.001). His bundle pacing also produced a greater acute hemodynamic response (4.6 mm Hg; 95% CI: 0.2 to 9.1 mm Hg; p = 0.04). The incremental activation time reduction with His bundle pacing over biventricular pacing correlated with the incremental hemodynamic improvement with His bundle pacing over biventricular pacing (R = 0.70; p = 0.04). Conclusions His resynchronization delivers better ventricular resynchronization, and greater improvement in hemodynamic parameters, than biventricular pacing.

L eft bundle branch block (LBBB) is deleterious in patients with heart failure. Biventricular pacing improves the activation pattern, instantly shortening left ventricular activation time (LVAT) and immediately improving cardiac function.
Long-term studies show substantial reduction in morbidity and mortality (1,2). Biventricular pacing (BVP) was first recognized as a potentially beneficial therapy from its clear hemodynamic effect (3). Relative to native LBBB, biventricular pacing (BVP) was found to shorten QRS duration (QRSd) and LVAT, earning it the moniker of "cardiac resynchronization therapy" (CRT). In fact, BVP results in the fusion of 2 nonphysiological wave fronts, resulting in only modest reductions in QRSd (4)(5)(6)(7). Computer modeling suggests that there is a large potential to deliver greater improvements in cardiac function if more effective ventricular resynchronization could be achieved (5).
His bundle pacing (HBP) has the potential to offer greater ventricular resynchronization because large reductions in QRSd have been observed when stimulating the His-Purkinje system in patients with LBBB (8,9). How this compares with BVP is beginning to be explored (9-11) but, crucially, there is no withinpatient comparison between HBP and BVP, of hemodynamic measurements alongside detailed electrical activation mapping.
In this prospective study, we tested the ability of HBP to deliver resynchronization. We then compared the electromechanical effects of His resynchronization against conventional BVP, using high-precision hemodynamic assessment (12) and noninvasive epicardial ventricular activation mapping (7).  Time (mm:ss) (Top left) A minimum of 4 alternations between AAI pacing and HBP or BVP were performed for each tested AV delay (therefore, a total of 8 transitions). For each alternation, relative change in SBP was calculated and the mean of the 8 transitions was calculated. (Top right) As more transitions occur, more data points are collected for change in SBP, eventually resulting in 10 values with mean and confidence intervals. Single or few measurements result in highly inaccurate estimates due to variability. (Bottom) A mean and confidence interval are calculated for each AV delay for both HBP (purple) and BVP (orange). An example is shown of quadratic curves fitted to data from multiple transitions at a range of AV delays to produce the peak systolic blood pressure responses. AV ¼ atrioventricular; BP ¼ blood pressure; BVP ¼ biventricular pacing; ECG ¼ electrocardiogram; HBP ¼ His bundle pacing; SBP ¼ systolic blood pressure.

METHODS
Arnold et al.   (18). Changes in this measurement correlate well with changes in invasively measured LV dP/dt (18), a measure of cardiac contractility that is largely independent of the degree of loading force.
The hemodynamic data were acquired alongside surface ECG data using a data acquisition system (National Instruments, Austin, Texas) and recorded using customized software. The 12-lead ECG recordings were made using the BARD electrophysiology lab system (Boston Scientific, Natick, Massachusetts) alongside ECGI. Values are mean AE SD (range) or n (%). The data are from 17 subjects in whom LVAT-95 was reduced by at least 10 ms and for whom data are available for both hemodynamic and electrical responses to His bundle and biventricular pacing.
ACE ¼ angiotensin-converting enzyme inhibit; ARB ¼ angiotensin receptor blocker; BP ¼ blood pressure; CS ¼ coronary sinus; LV ¼ left ventricular; MI ¼ myocardial infarction; MRA ¼ mineralocorticoid receptor antagonist; NYHA ¼ New York Heart Association.   ANALYSIS OF HEMODYNAMIC DATA. The peak blood pressure response and its CIs were calculated by fitting a quadratic curve to the data from each set of tested AV delays ( Figure 1). Analysis of hemodynamic data was automated using Python, version 3.6 (Python Software Foundation, Wilmington, Delaware) so that few user inputs were required. Robust regression methods were used so that outliers did not need to be manually removed

RESULTS
Twenty-three patients were recruited. In 4 patients, temporary HBP did not shorten LVAT-95 by $10 ms, and these patients were therefore excluded from the main analysis. In 1 patient, a technical fault prevented acquisition of ECGI data, and they were therefore also excluded from the analysis.    Figure 4). An example of epicardial activation maps and ECGs are shown in Figure 5. p < 0.001, respectively) in the 18 eligible patients in whom the full hemodynamic dataset was acquired.

RELATIONSHIP BETWEEN ELECTRICAL AND
HEMODYNAMIC RESPONSE. There was a significant correlation (R 2 ¼ 0.49, R ¼ 0.70; p ¼ 0.04) (Figure 7) between the individual patient incremental    His bundle pacing delivered significantly greater within-patient improvements in acute systolic blood pressure (right). The 95% confidence intervals are displayed. Abbreviations as in Figures 1 and 2.
Arnold et al.   Although noninvasive ECGI mapping provides high-resolution data of activation and allows segregation of regional activation, there are potential problems including assumptions of static geometry and interpolation accuracy (31) (Online Appendix).
We did not observe successful ventricular resynchronization with HBP in all patients in this study, which is consistent with the findings of other studies (8). As this was an acute hemodynamic study of temporary HBP, we did not employ active fixation, which is used in permanent HBP. Active fixation may result in lower thresholds because the screw of the active lead acts as the cathode, and therefore, when it penetrates the His bundle, it may allow direct stimulation of the conduction fibers. However, some patients with QRS prolongation may not be amenable to ventricular resynchronization with His pacing, for example, patients where QRS prolongation occurs due to myocyte uncoupling.

CONCLUSIONS
CRT delivered using HBP appears to be a very promising alternative to BVP in patients with LBBB and heart failure. It can deliver larger reductions in ventricular activation time, which leads to significantly greater improvements in acute hemodynamic function. The magnitude of these improvements suggest that His resynchronization therapy has the potential to produce better clinical outcomes than BVP.
ACKNOWLEDGMENT Dr. Ben Ariff, Imperial College London, oversaw the radiological support for Car-dioInsight CT imaging. His Pacing Versus Biventricular Pacing for LBBB