Introduction

Cardiac resynchronization therapy (CRT) has been shown to improve symptoms and survival in patients with systolic dysfunction and prolonged ventricular depolarization (wide QRS).1-8 However, atrial fibrillation (AF) may cause significant problems in these patients, particularly when the ventricular rate exceeds the device programmable rate. Atrial fibrillation in heart failurepatients is often clinically challenging,9 and if poorly controlled, permanent atrioventricular (AV) nodal ablation when pharmacologic rate control is not possible may be required.10 For the past 5 years, our laboratory has been studying the potential application of coupled pacing (CP) as a means of rate control when AF occurs.11 - 14 Briefly, CP can be explained as follows: after sensing the intrinsic electrical activation of the ventricles which initiate mechanical contraction, an additional electrical stimulation (CP) is applied near the end of the T-wave. Coupled pacing is applied prior to the time that the ventricles are capable of fully contracting again. These critically timed retrograde activations of the AV node in turn prevent subsequent rapid ventricular activations that would have lead to weakened ventricular contractions. An additional advantage of CP in patients with AF and systolic dysfunction is that the CP beat increases contractility via the mechanism of postextrasystolic potentiation.15 Since premature stimulations have differential effects on Purkinje vs. ventricular muscle refractoriness, 16,17 there is some concern that CP may potentially alter the left ventricular (LV) electrical activation pattern during biventricular pacing. The goal of our study was therefore to better define the effects of CP during biventricular pacing (CRT + CP) on LV function and mechanical dyssynchrony.

Methods

The experimental protocol was approved by the Animal Research Committee of the Cleveland Clinic. All animals received humane care in compliance with the 'Guide for the Care and Use of Laboratory Animals'. The dogs (n ¼ 6) were initially anesthetized with thiopental and maintained with isoflurane during positive pressure ventilation. A mid-sternotomy was performed and the heart was placed in a pericardial cradle. A quadrapolar plate electrode was sutured to the right atrium and was connected to a Grass stimulator for the induction of a pseudo AF. A second quadrapolar electrode was sutured on the right ventricular (RV) apex and was connected to the first channel of a Bloom stimulator for RV pacing as the protocol dictates. A third electrode was sutured on the lateral wall of the LV and connected to the second channel of our Bloom stimulator for LV pacing. The remaining two poles of these three quadrapolar electrodes were connected to our PonemahTM system (Valley View, OH, USA) to record their corresponding electrograms: right atrial electrogram (RAE), right ventricular electrogram (RVE), and left ventricular electrogram (LVE). Finally, a bipolar plate electrode was sutured on the inferior vena cava-left atrial epicardial fat pad to stimulate the parasympathetic nerve that innervates the AV node to slow A-V nodal conduction.18

Pacing protocol

Five to ten minutes of each pacing paradigm was applied before obtaining the echocardiographic and invasive haemodynamic acquisitions. Step 1: With the animal in sinus rhythm (SR). Step 2: Subsequently, a simulated paroxysmal AF was induced by continuous rapid right atrial pacing (20 Hz, 1 ms, 3-5 V). Importantly, rapid atrial pacing was maintained continuously to ensure persistent AF for Steps 3-6.11 - 13 Step 3: Application of rapid RV pacing (2 ms, 2-4 mA) at a ventricular rate greater than the resultant ventricular rate from rapid atrial pacing alone. The purpose of this rapid RV pacing was to prevent intrinsic ventricular activation over the AV node, thereby inducing dyssynchrony comparable with left bundle branch block (LBBB). Step 4: CRT: Rapid simultaneous biventricular pacing (both channels of the Bloom stimulator set at 2 ms, 2-4 mA) allowing capture by both ventricles in the presence of rapid simulated AF. Step 5: CRT + CP: Simultaneous biventricular pacing (both channels of the Bloom stimulator), followed by an additional stimulation of CP (first channel of the Bloom stimulator) which was applied only to the RV lead near the end of the T-wave. This was done by increasing the basic cycle length during biventricular pacing by 50% above the length used in Step 4 (CRT alone). We subsequently added CP stimuli at a delay of 250 ms. This delay was then progressively shortened until we observed by both echocardiographic and LV pressure recordings that CP resulted in only minimal LV mechanical contractions (not leading to effective ejection). By adding CP, we were able to finally reduce the biventricular pacing rate to a rate close to SR. Step 6: vagal stimulation (CRT-VS): CRT capture enabled by reducing the ventricular rate response during rapid AF with selective stimulation of the parasympathetic nerves which innervate the AV node (pulses ¼ 20 Hz, 0.1 ms, 10 mA). At this level of intensity, this selective and limited stimulation of the preganglionic parasympathetic fibres projection towards the AV node18 slowed the rate of ventricular activation from the atria without negative inotropic effect. This allowed applying the CRT at a rate also similar to the SR. In this study, we chose not to apply CP to the LV for the following reasons: (1) The timing of retrograde activation of the AV node/His system would have been difficult to control if left ventricular CP had been applied in the presence of intrinsic LBBB. However, LBBB was not present in our study. (2) Since scar tissue would most likely be in the LV, applying CP close to scar tissue may theoretically increase the risk of fatal arrhythmias.19 Thus for these reasons, we chose to evaluate the effects of CP when it was applied only from the RV.

Epicardial echocardiographic data

Echocardiographic acquisitions were performed with a Vivid 7 machine (GE Healthcare). During these periods, we turned off the aortic flow meter in order not to interfere with the Doppler measurements from the echocardiograph. The time intervals between two Doppler LV outflow peaks were used to determine the cardiac period. These integrated velocity profiles, times aortic cross-sectional area, were used to obtain an estimate of stroke volume (SV). Left ventricular contractile function was quantified by measuring (1) left ventricular ejection fraction (LVEF) and (2) the peak circumferential global LV strain by speckle tracking. Left ventricular ejection fraction was computed from standard apical views by using Simpson's biplane method. Global circumferential strain curve was derived from the 2D short axis view (basal segment) using 2D strain analysis (EchoPac, GE Healthcare). Peak circumferential global strain was defined as the minimal strain value during cardiac cycle. Left ventricular dyssynchrony was calculated as the standard deviation (SD) of the time (%) to peak circumferential strain of these six different segments with a larger SD indicating an increasing degree of dyssynchrony. Diastolic filling time [duration of the early diastolic filling wave (E) during AF or the sum of E and late diastolic filling wave (A) during SR] was normalized to cardiac cycle for variability in cycle length. All echocardiography parameters used were obtained by averaging five consecutive values.

. . .Continue to read rest of article (PDF).