TEXT
Almost 50 years ago, one of us (Bourassa MG) co-authored a case report entitled: “Hemodynamic Studies during Intermittent Left Bundle Branch Block”[1]. The patient was a relatively young man with symptoms and signs of heart failure and a heart murmur suggestive of aortic regurgitation. Intermittent left bundle branch block (LBBB) was documented on electrocardiograms (ECG), vectorcardiograms and phonocardiograms prior to cardiac catheterization. During left heart catheterization, the ECG spontaneously changed from a pattern of left ventricular (LV) hypertrophy (QRS duration: 80 ms), to a pattern of LBBB (QRS duration: 160 ms). Following oxygen administration, the ECG rapidly reverted to a pattern of LV hypertrophy with normal conduction. During the procedure, similar ECG sequences occurred during which peripheral, aortic and LV pressure curves were simultaneously recorded and cardiac output was calculated using dye-dilution curves.
Important hemodynamic changes consistently accompanied the occurrence of intermittent LBBB. Notably, systolic pressures in the left ventricle, central aorta and radial artery fell consistently (roughly 20 mmHg during LBBB as compared to normal conduction); cardiac index fell from 2.2 L/min per m2 during normal conduction to 1.7 L/min per m2 during LBBB. The ECG showed an immediate increase in heart rate during LBBB, presumably as an attempt to compensate for the decreased cardiac output. Particularly important was the temporal relationship of cardiac events in both forms of conduction. The time interval between onset of LV depolarization and onset of isometric contraction remained unchanged, and thus, the onset of isometric contraction was not delayed during LBBB. On the other hand, slowing of intraventricular conduction appreciably prolonged the duration of isometric contraction (from 72 ms during normal conduction to 94 ms during LBBB) and, proportionately, of isometric relaxation. Because of prolongation of isometric contraction, both the onset and termination of systolic ejection were delayed during LBBB, although the duration of systolic ejection itself was unchanged. The diastolic period with LBBB was shorter than with normal conduction.
To our knowledge, this observation provided the first ever evidence that slowing of intraventricular conduction and prolonged isometric LV contraction result in decreased force and efficiency of contraction, leading to a notable reduction in cardiac output, a drop in systemic blood pressure and a compensatory increase in heart rate. Conversely, spontaneous or oxygen-induced (in our case) conversion from LBBB to normal LV conduction promptly returned hemodynamic parameters to their previous levels. Thus this severe LV dysfunction was immediately reversible.
LBBB results in a significant delay in aortic opening and closure, but it does not affect the timing of RV events[2,3]. It has been suggested that interventricular dyssynchrony contributes to reduction in the regional ejection fraction (EF) of the septum without impacting LV apical and lateral wall motion[4]. On the other hand, intraventricular dyssynchrony, which is characterized by heterogeneous activation of different LV segments (some being activated early and others late during cardiac contraction), results in decreased cardiac output, systemic blood pressure, maximal rate of pressure rise (dP/dt), and global EF[3,5,6]. Finally, in patients with sinus rhythm, atrial contraction is not followed by a properly timed LV systole and prolonged atrioventricular delay can also contribute to cardiac dysfunction.
In the mid-1990s, some investigators hypothesized that patients with LV dysfunction and delayed intraventricular conduction would benefit from pacing at sites that achieve a more rapid ventricular depolarization and thus a more synchronous contraction[7,8]. This led to evaluation of atrial-synchronized biventricular pacing as a means to resynchronize ventricular contraction and improve cardiac function[7,8]. Cardiac resynchronization therapy (CRT) was shown to improve LVEF and dP/dt, and to reduce LV end-diastolic and end-systolic volumes[9]. Following quite convincing observational data, several large randomized clinical trials demonstrated that multisite ventricular pacing or CRT significantly improved mortality and morbidity in patients with heart failure and complete LBBB[10-14].
Currently, CRT is an established treatment modality for selected patients with systolic heart failure. Selection criteria for implantation of a CRT device include an EF < 35%, NYHA functional class III-IV symptoms, and a wide QRS complex (duration ≥ 120 ms). In patients with mild heart failure, the RAFT study found that CRT provided additional benefit to an implantable defibrillator if intrinsic QRS duration was 150 ms or more[14]. This ECG criterion was incorporated into European guidelines for CRT patients with NYHA class II symptoms[15]. Thus, indications continue to expand to less symptomatic patients[14]. However, the intervention is both invasive and costly, and clinical response and long-term outcome are variable. Roughly 30% of patients who meet established criteria do not respond to CRT[7-13,15]. In patients with ischemic cardiomyopathy, Barsheshet et al[16] found that those with a QRS duration ≥ 150 ms, blood pressure < 115 mmHg, or LBBB had a more favorable response to CRT with regards to overall mortality or heart failure events. Thus, the subgroup of patients at higher risk for death or heart failure appeared to derive the greatest benefit from CRT. In contrast, identified predictors of a favorable response to CRT in non-ischemic cardiomyopathy were female gender, diabetes, and LBBB[16]. While further studies are required to elucidate pathophysiological mechanisms and clinical implications of such subgroup analyses, current evidence suggests that patients with a wide QRS and RBBB or non-specific intraventricular conduction abnormalities respond less favorably to CRT regardless of the type of cardiomyopathy[17,18].
Predictors of a favorable outcome after CRT include baseline mechanical LV dyssynchrony, optimal LV lead position, and extent and location of myocardial scar[19,20]. However, the pathophysiology of mechanical LV dyssynchrony is complex and both its quantification and the choice of optimal, concordant pacing sites, in the latest mechanically activated region of the LV and remote from an infarct zone, still remain under investigation. The Predictors of Response to CRT trial revealed the technical problems of echocardiographic parameters, such as those derived from M-mode echo, routine pulsed Doppler and tissue Doppler imaging, for predicting response in patients with standard CRT indications[21,22]. More recently, promising new tools such as 2-D speckle tracking and real-time 3D echo, alone or in association with cardiac magnetic resonance imaging, have been shown to improve patient selection for CRT[22]. Other investigators have assessed new lead placement strategies and have concluded that CRT delivered at the optimal LV endocardial sites (sites of latest mechanical activation) is more effective than via coronary sinus lead pacing[19,20,23]. Future investigations of epicardial or LV endocardial lead positioning may eventually overcome current anatomical restraints[23].
Finally, it has been shown recently that evidence of mechanical LV dyssynchrony at baseline, usually associated with increased QRS duration, and definite reduction in LV dyssynchrony post-implantation are essential in predicting a positive response to CRT[24]. Our seminal report clearly illustrated the importance of these two mechanisms. Delayed intraventricular contraction was associated with deterioration, and return to a normal pre-ejection phase was associated with improvement in LV function.
