INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

HEART FAILURE (HF) affects more than 5,000,000 Americans, with more than 400,000 new cases diagnosed each year. A significant percentage of HF patients presents with an abnormally wide QRS complex, often resulting from complete or partial left fascicular block. Recent studies (3-5, 7, 8, 10, 12, 16, 22) have shown that cardiac resynchronization therapy (CRT) with biventricular (BV) or left ventricular (LV) stimulation in the New York Heart Assocation functional class III and IV HF patients with wide QRS and left bundle branch block (LBBB) can improve acute hemodynamic function and chronic functional status. Moreover, the acute improvements in systolic function seem to be accompanied by decreased energy consumption in the myocardium (21). It has been reported (4, 10, 16) that a wide QRS complex and the presence of LBBB may be useful clinical markers to identify those patients who may benefit from CRT.

The mechanisms suggested for the hemodynamic improvements observed with CRT include restoration of optimal atrioventricular (AV) timing and septal-LV lateral-posterior wall recoordination by LV or BV stimulation. However, the isolated role of LBBB in the acute hemodynamic improvements observed during CRT has not been previously studied. The aim of this study, therefore, is to develop an animal model of isolated LBBB, which increases the width of the QRS complex by delaying the activation of the LV lateral free wall. Second, we sought to use this LBBB model to test the hemodynamic impact of this electrical delay and whether CRT would restore cardiac function by correcting the asynchronous contraction of the LV.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study was conducted in nine adult greyhound dogs of either sex weighing 29.9 ± 3.9 kg. The stimulation protocol was performed in six dogs to obtain simultaneous electrophysiological and hemodynamic data. Echocardiography was performed in three more dogs to determine the relative coordination of the septal and LV free wall motion. Dogs were premedicated with 10 mg of butorphanol and 0.5 mg of acepromazine administered subcutaneously. General anesthesia was induced with 150 mg of ketamine and 7.5 mg of diazepam, and was maintained with the use of isoflurane gas and a semi-pressure-volume-regulated ventilator. Heart rate and blood pressure were monitored during the procedure to ensure that a deep level of anesthesia was maintained.

All experiments were carried out in accordance with National Institutes of Health's Guide for the Care and Use of Laboratory Animals. The Institutional Animal Care and Use Committee of Guidant Corporation approved this protocol.

Catheterization. After the jugular veins and the right carotid artery were exposed, a 7-Fr pulmonary-wedge pressure catheter (Bard; Billerica, MA) was advanced into the ostium of the coronary sinus under fluoroscopic guidance. A coronary venogram was generated by inflation of the balloon and injecting renographin into the coronary sinus. The image was then stored and displayed on a monitor to guide stimulation lead placement. After the balloon catheter was removed, two 8-Fr custom-designed guiding catheters were advanced into the ostium of the coronary sinus, which served as conduits for the coronary vein leads. Two coronary vein prototype leads were implanted at the apex and base of the LV in anterior and lateral branches of the great cardiac vein, as described by Auricchio et al. (2). A unipolar lead (Sweet Tip, model 4169, Guidant; St. Paul, MN) and a bipolar lead (model 4269, Guidant) were placed in the right ventricular (RV) apex and right atrium (RA) for stimulation and sensing, respectively. Two 8-Fr dual-transducer pressure catheters (model SPC-780c, Millar Instruments; Houston, TX) were placed in RV and LV to measure RV, LV, and aortic pressures (Fig. 1). Pressure catheters and stimulation leads were connected to a custom external stimulation system (FlexStim, Guidant) to acquire hemodynamic signals and execute an acute stimulation protocol.

View larger version (138K): [in this window] [in a new window]   Fig. 1.   Fluoroscopic image of leads and catheter placement. RA, right atrial bipolar lead; RV, right ventricular unipolar lead; LV1, left ventricular (LV) basal unipolar coronary vein lead; LV2, LV apical unipolar coronary vein lead; LVP, LV pressure transducer; AoP, aortic pressure transducer; RVP, RV pressure transducer.

Creation of LBBB. To generate LBBB, an 8-Fr-introducing sheath was inserted into the carotid artery. A 7-Fr ablation catheter with a 4-mm tip (model D7-DL-252-PS, Cordis Webster; Baldwin Park, CA) was then advanced retrogradely across the aortic valve into the LV and manipulated under fluoroscopic guidance until a discrete left bundle potential (LBP) was found in the distal bipole and recorded via a multichannel oscilloscope (Fig. 2). The LBP was confirmed by an atrial and ventricular amplitude ratio <1:10 and an interval of <35 ms from LBP to QRS complex. LBB ablation was performed with a radio frequency (RF) generator (model RFG 3-E, Radionics; Burlington, MA) delivering 480 kHz (±10%) unmodulated sine wave energy. RF energy of 30 W was applied for 60 s between the tip of an ablation catheter and a metal plate placed on the back of each dog (15). The impedance was continuously monitored during the energy application. A drop in impedance was considered a sign of good tissue contact and adequate heating (14). The energy application was continued for 60 s or until a sudden rise in impedance occurred, which indicated that coagulation had formed at the tip of the catheter.

View larger version (73K): [in this window] [in a new window]   Fig. 2.   Electrocardiogram (ECG) surface lead II and an electrogram recorded from the radio frequency (RF) ablation catheter. Top trace, surface ECG lead II; middle trace, intracardiac electrogram recorded from the distal bipole of the RF ablation catheter; bottom trace, intracardiac electrogram recorded from the proximal bipole of the RF ablation catheter. A, atrial signal; V, ventricular signal; LBP, left bundle potential; T, T wave on surface ECG; P, P wave on surface ECG; QRS, QRS complex on surface ECG. Paper recording speed was 250 mm/s.

Stimulation protocol. Before ablation and 30 min after successful LBB ablation, a software-controlled stimulation system (FlexStim, Guidant) was connected to the animal and an acute stimulation protocol was executed. The protocol was designed to measure the immediate hemodynamic effects of CRT while accounting for local baseline shifts. This allowed statistical comparison of multiple stimulation combinations within individuals as described previously (4). Briefly, RV, LV, or BV stimulation was performed in a VDD mode (i.e., atrial sense followed by ventricular stimulation) for six beats after 14 sinus beats at 1 of 4 preset AV delays. The AV delays were determined by equally dividing the interval between 8 ms and the intrinsic AV interval 30 ms into four parts. Each combination of stimulation chamber and AV delay was repeated four times in a random order. Intracardiac electrogram (EGM) and pressure signals were simultaneously recorded to the computer hard disk for off-line analysis. Off-line analysis was performed with custom software that automatically calculated the following: aortic diastolic pressure (ADP), aortic systolic pressure (ASP), pulse pressure (PP = ASP  ADP), LV maximum pressure derivative over time (LV dP/dt_max), minimum dP/dt (LV dP/dt), and LV end-diastolic pressure (LVEDP).

Echocardiography. Because positive hemodynamic response during LV or BV stimulation after LBB ablation was observed in all six dogs, echocardiography was adopted to qualitatively look at the LV and septal wall motion to prove the hypothesis that preexcitation of LV in the presence of LBBB improves cardiac function by coordinating the LV contraction. M-mode echocardiograms (Sonos 2500, Hewlett-Packard) were obtained in three dogs by placing a 2.5-MHz transducer on the right side of the canine thorax. The probe was positioned to track the interventricular septum and LV posterior wall motion, at the level of the tips of the papillary muscles. Echo recordings were obtained during normal sinus rhythm at baseline, after LBB ablation, and during LV stimulation at an optimum AV delay selected by the FlexStim stimulation protocol in the presence of LBBB. Because LV and BV stimulation improved hemodynamics to a similar extent, LV stimulation was applied during Echo examination in all three dogs. Echocardiograms from each experiment were recorded on videocassette for off-line review.

Gross necropsy. After each experiment, the heart was excised and rinsed with saline. The left ventricle was carefully dissected and the endocardial surface was wiped with Lugol's solution to expose the structure of the conduction system (27). Each necropsy was archived by photography and the location of all ablation lesions was determined by visual inspection.

Statistical analysis. Hemodynamic response to stimulation was determined by following a previously published method (4). The percentage change for a given hemodynamic parameter was calculated from the value of the parameter during stimulation compared with the average value during the immediately preceding six nonpaced beats (i.e., local baseline) for each tested AV delay. The first two paced beats were ignored in the analysis because their hemodynamic response is biased by the diastolic behavior of the preceding nonpaced beat and the transitory ventricular V-to-V decrease caused by the switch to a short AV delay (4). In addition, sequences containing ectopic beats were automatically repeated by the system and ignored during the off-line processing phase. A two-way ANOVA was applied to analyze differences between stimulation chambers (RV, LV, and BV) before and after LBB ablation. The AV delay, LBBB, and stimulation chamber were considered the treatment variables. A paired t-test was used to compare the effects of two group data. A P value of <0.05 was considered statistically significant. Average data are shown as means ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Creation of LBBB. A median of one RF ablation (range 1-3) was performed to create block of LBB in nine dogs. A successful ablation was confirmed by prolonged QRS duration (range 123-160 ms), LBBB morphology (see the description in METHODS), and a constant PR interval on the surface ECG (Fig. 3). The mean interval from LBP to ventricular electrogram at the successful ablation site was 33.4 ± 2.5 ms (range 29-35 ms). Successful LBB ablation doubled QRS duration (P < 0.001) and shifted QRS axis (P < 0.05) with no substantial changes in HR and PR interval (Table 1).

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Representative 12-lead surface ECG before and after left bundle branch (LBB) ablation.

Interventricular synchrony. Table 2 reports changes of interventricular delay as defined by either the interval between RV and LV deflection on the EGM or the onset of systole on the RV and LV pressure curves for preablation (baseline), postablation (LBBB), and during BV or LV stimulation. The value of LV pressure onset during stimulation reported in Table 2 was taken from the stimulation configuration (either LV or BV) with the percentage of LV dP/dt_max response in each dog. At baseline (sinus rhythm with intact conduction system), RV and LV started to activate and contract almost simultaneously (~1 ms). After the LBB was ablated, however, both LV activation and contraction were significantly delayed (~30 ms) relative to RV (P < 0.001 for both). In the presence of LBBB, either BV or LV stimulation significantly shortened RV-LV contraction delay and improved interventricular asynchrony (P < 0.001 vs. LBBB), although it did not completely restore the normal contraction sequence (P < 0.05 vs. baseline).

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Effect of pacing site and atrioventricular (AV) delay on LV maximum pressure derivative over time (dP/dt) (A) and pulse pressure (PP) (B) before and after LBBB.

Echocardiography. Figure 5 shows the M-mode echocardiograms of the LV at baseline (Fig. 5A), after LBB ablation (Fig. 5B), and during LV stimulation in the presence of LBBB (Fig. 5C). During intrinsic rhythm before LBBB (Fig. 5A), the interventricular septum moved towards the LV posterior wall during ventricular contraction whereas the posterior wall of the LV moved anteriorly towards the septum. The simultaneous inward movement of both the septum and free wall efficiently ejected the blood from the LV into the aorta. In contrast, LBBB delayed LV activation (Table 2) and caused the septum to move downward after the onset of electrical depolarization (vertical line in Fig. 5B), followed by movement away from the LV posterior wall at the time when the LV started to contract. This motion produced a downward beaking (arrow in Fig. 5B) of the left septum shortly after depolarization and a parallel movement of the interventricular septum and the LV posterior wall during ventricular ejection. Such a paradoxical septal motion caused inefficient propelling of blood from LV chamber into the aorta. However, in the LBBB model, LV stimulation with a shortened AV delay produced M-mode traces (Fig. 5C) similar to those seen in normal sinus rhythm.

View larger version (86K): [in this window] [in a new window]   Fig. 5.   M-mode echocardiograms of the left ventricle at baseline (A), after LBBB (B), and during LV stimulation in the presence of LBBB (C). RV, right ventricular chamber; IVS, interventricular septum; LV, left ventricular chamber; LVPW, posterior wall of the left ventricle. Vertical line marks the beginning of QRS complex.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, we have developed an animal model of LBBB using focal endocardial RF ablation. LBBB resulted in delayed LV activation and a corresponding delayed LV systole. This phenomenon was associated with an asynchronous contraction of the septum and the LV free wall. This asynchronous activation and contraction resulted in decreased LV global function as evaluated by PP and LV dP/dt_max, respectively. Because PP and stroke volume changes have been proven to be correlated with each other during both steady-state pacing in a CHF patient population (16) and burst pacing using FlexStim protocol in an animal study (19), and provided that LBB ablation did not affect HR (Table 1), the increase in LVEDP and decrease in PP was most likely caused by a less efficient LV pump and associated with a lower cardiac output. Although ventricular stimulation in normal dogs worsened global ventricular function, LV or BV stimulation significantly improved LV systolic and diastolic function in canines with LBBB, probably by correcting the interventricular contraction asynchrony. These findings support the hypothesis that LV and BV stimulation provide hemodynamic benefit by improving electromechanical coordination of LV contraction in the presence of LBBB.

This hypothesis is further supported by the fact that the improvement in global LV function as evaluated by dP/dt_max obtained with LV and BV stimulation occurred at every AV delay from 10 to 100 ms (Fig. 4A). PP, in contrast to LV dP/dt_max, showed a marked decrease when the AV delay was shortened <70 ms (Fig. 4B) which indicated that preload was an important factor for PP. The independence of the paced AV delay and the magnitude of the LV dP/dt_max improvement may indicate that neither left-sided AV synchrony nor preload is the primary mechanism mediating LV dP/dt_max improvements provided by LV and BV stimulation in this model.

Canine LBBB model. LBB ablation produced interventricular asynchrony and deteriorated hemodynamics in the normal canine heart. During sinus rhythm, the onset of the LV electrogram and the upstroke of the LV pressure occurred almost simultaneously with the RV, and the LV pressure exceeded the RV pressure throughout the cardiac cycle. The echocardiogram showed that the interventricular septum moved posteriorly as it began to contract (Fig. 5A). This coordinated contraction pattern was immediately lost with LBBB (Table 2). In the LBBB heart, the LV free wall was in its relaxation phase while the RV started to contract, causing an abrupt posterior motion of the interventricular septum. The delayed onset of LV contraction, occurring as the septum begins to relax, resulted in a paradoxical septal movement whereby the septum moved away from the LV posterior wall during LV systole (Fig. 5B: Beak) and a decreased septal contribution to stroke volume.

Comparison of canine LBBB model to human LBBB. The asynchronous behavior and the hemodynamic changes caused by LBB ablation in our model are similar to the abnormalities found in LBBB patients. Abnormal interventricular septal motion in patients with LBBB has been described since the 1970s (1, 9, 20). Abbasi et al. (1) reported that 14 of 17 patients with complete LBBB had abnormal interventricular septal motion analogous to that observed in the present study. Furthermore, in 2 of 14 cases with intermittent LBBB, abnormal septal motion was present only during LBBB. This abnormality was explained by asynchronous LV contraction with early activation and contraction of the septum but delayed activation and contraction of the LV free wall (20).

Clinical relevance. As early as the 1970s, researchers (11) proposed that LBBB could be a predisposing factor for subsequent congestive cardiomyopathy; however, it was believed that congestive cardiomyopathy would develop only with additional specific influences, such as arterial hypertension, viral infection, alcohol, or pregnancy. Later, Kuhn et al. (17) observed that latent cardiomyopathy might be an advanced stage for patients with lone LBBB. Furthermore, in the Framingham Heart study, people with acquired bundle branch block, particularly LBBB, were more likely to develop advanced cardiovascular disease (23). Several studies (24, 25) have also found increased mortality in patients with LBBB, and the effect of LBBB on mortality is most pronounced in those with severe LV dysfunction.

Limitations. The LBBB model created in the present study was discrete and may not mimic diffuse changes of the conduction system due to progressive chronic pathological changes in HF patients with dilated cardiomyopathy. This LBBB model was developed in normal hearts that exhibit no other pathophysiological alterations such as myocardial fibrosis, necrosis, or apoptosis. Results from experimental models or humans with more complex substrates may differ from those of the present study. However, our model was intended to isolate the conduction delay component observed in many HF patients without the confounding effects of other pathophysiological substrate changes. The model is intended to serve as a building block for the inclusion of more complex myocardial insults to more closely match the clinical condition. Regardless, the hemodynamic improvement observed during CRT in this study was quite similar to that observed in HF patients with various etiologies and extent of conduction delay.