OUP user menu

Greater response to cardiac resynchronization therapy in patients with true complete left bundle branch block: a PREDICT substudy

Mark Jonathan Perrin, Martin S. Green, Calum J. Redpath, Pablo B. Nery, Arieh Keren, Robert S. Beanlands, David H. Birnie
DOI: http://dx.doi.org/10.1093/europace/eur381 690-695 First published online: 14 December 2011


Aims Cardiac resynchronization therapy (CRT) benefits patients with heart failure and a wide QRS complex. Still, one-third derive no clinical benefit and a majority of patients demonstrate no objective improvement of left ventricular (LV) function. Left bundle branch block (LBBB) is a strong predictor of response to CRT. We evaluated whether absence of electrocardiogram (ECG) markers of residual left bundle (LB) conduction in guideline-defined LBBB predicted a greater response to CRT.

Methods and results An r wave ≥1 mm in lead V1 (r-V1) and/or a q wave ≥1 mm in lead aVL (q-aVL) was used to identify patients with residual LB conduction. Forty patients with a wide QRS were prospectively enrolled and subdivided into three groups: complete LBBB (cLBBB), LBBB without r-V1 or q-aVL (n = 12); LBBB with residual LB conduction (rLBBB), LBBB with r-V1 and/or q-aVL (n = 15); and non-specific intraventricular conduction delay (IVCD), (n = 13). Following CRT: mean change in left ventricular ejection fraction was 11.9 ± 11.9% in cLBBB, 3.8 ± 5.4% in rLBBB (P= 0.045), and 2.5 ± 4.4% in IVCD (P= 0.02 cLBBB vs. IVCD); mean reduction in left ventricular end-systolic volume was 26.4 ± 39.2% in cLBBB, 14.3 ± 22.9% in rLBBB (P= 0.35), and 5.6 ± 17.3% in IVCD (P= 0.11 cLBBB vs. IVCD); mean change in native QRS duration was −8.0 ± 11.0 ms in cLBBB, −0.8 ± 8.24 ms in rLBBB (P= 0.07), and 0.15 ± 8.0 ms in IVCD (P= 0.048 cLBBB vs. IVCD).

Conclusion In patients with guideline-defined LBBB, the absence of ECG markers of residual LB conduction was predictive of a greater improvement in LV function with CRT.

  • Cardiac resynchronization therapy
  • Electrocardiogram
  • Electrophysiology
  • Left bundle branch
  • Cardiac failure



Cardiac resynchronization therapy (CRT) reduces heart failure symptoms,13 improves left ventricular (LV) function, and decreases mortality in New York Heart Association (NYHA) II–NYHA IV (ambulatory) heart failure with a wide QRS complex.3,4 Still, ∼30% of patients derive no clinical benefit with CRT5,6 and a smaller percentage display reverse LV remodelling, or improved LV function.7,8 Generally speaking, the magnitude of improvement in LV ejection fraction (LVEF) and reduction in LV dimensions in patients with heart failure is proportional to survival;9,10 specifically, reduction in LV dimension but not clinical response to CRT is associated with reduced mortality.8 Conversely, CRT without objective LV improvement may be proarrhythmic.11

The 12-lead electrocardiogram (ECG) remains one of the strongest predictors of response to CRT: a QRS duration of ≥150 ms pre-implant;3,12,13 a reduced QRS duration (delta-QRS) with LV pacing;6,14 and a left bundle branch block (LBBB) morphology on the pre-implant ECG—patients with right bundle branch block (RBBB) or non-specific intraventricular conduction delay (IVCD) showing little or no benefit.1518

An ECG showing LV conduction delay may be due to: true proximal LBBB; delay in the distal conduction system; or LV conduction delay. Initial left-to-right septal forces should be absent in complete LBBB (cLBBB) (Figure 1). Thus, current guidelines of LBBB require absence of q waves in lead I and V5/V6.19 However, an r wave in V1 or q wave in aVL is also suggestive of left-to-right septal activation though both are allowable in current guidelines defining LBBB.19 Padanilam et al.20 recently reported that in the context of ‘LBBB’ an r wave ≥1 mm in V1 (r-V1) was predictive of residual left bundle (LB) conduction; a single patient in this study with a q wave in aVL >1 mm (q-aVL) had a similar result. Therefore, these simple ECG markers may define a form of IVCD that is overlooked by current guidelines.

Figure 1

Hypothesized early ventricular activation in complete left bundle branch block and ‘left bundle branch block’ with residual left bundle conduction. Note that V1 is in close proximity and near perpendicular to the interventricular septum. Lead V6, I, and aVL are approximately in the same coronal plane but en face to different surfaces of the heart—V6 lower than I and aVL. In complete left bundle branch block the septum is activated from right to left (B) and opposed by endocardial to epicardial activation of the right ventricular free wall (A); this composite activation results in a q wave in V1 and/or an r wave <1 mm. When residual left-sided conduction remains, septal activation is left to right (C) and summed with right ventricular free wall activation (A) resulting in an r wave in V1 ≥1 mm.

The data obtained from the PREDICT study21 were analysed to test our hypothesis that patients with true cLBBB would have a greater response to CRT compared with patients with ‘LBBB’ and residual LB conduction. For this purpose, we prospectively defined an r in V1 ≥1 mm (r-V1) and/or a q in aVL ≥1 mm (q-aVL) as markers of residual LB conduction in the context of guideline-defined LBBB.



The PREDICT study21 prospectively enrolled consecutive consenting patients with LVEF ≤ 35% and NYHA class II or III symptoms on optimal medical therapy with a QRS duration of ≥130 ms. Patients with RBBB, atrial fibrillation, or with in situ pacemakers were enrolled as part of PREDICT but excluded from this substudy. Patients were followed for 3 months post-implantation.

Ischaemic aetiology was defined as a history of myocardial infarction, or evidence of significant coronary artery disease on coronary angiography (stenosis ≥70% in two major arteries). Non-ischaemic aetiology was identified by the absence of documented myocardial infarction and no significant coronary artery disease on angiography (no stenosis ≥50%).

The following investigations were obtained before CRT implantation and analysed for this substudy: 12-lead ECG, positron emission tomography (PET) perfusion and metabolism imaging, using rubidium-82 for perfusion and F-18-fluorodeoxyglucose for metabolism, and an equilibrium radionuclide ventriculogram (ERVG). All studies were repeated, with the exception of the PET scan, 3 months post-implantation.

At 3 months post-implantation, CRT was turned-off briefly to compare the native QRS duration to that pre-implant (delta native QRS).


The procedures for data acquisition and analysis for: single-photon emission computed tomography-ERVG imaging and the quantification of mechanical dyssynchrony have been previously described.21 In brief, in-house software was used to create 568 radial profiles for phase analysis.22 The program assigns a phase angle to each pixel of the phase image, derived from the first Fourier harmonic of the time-activity curve for that pixel. The standard deviation of the phase angles is a measure of the extent of LV mechanical dyssynchrony.22 Phase analysis has been shown to have good correlation with tissue Doppler echo techniques for assessing dyssynchrony.22 Equilibrium radionuclide ventriculogram was employed to measure LVEF.

Positron emission tomography imaging and analysis, and tissue characterization, were performed as previously described.21

Cardiac resynchronization therapy implantation and optimization

The atrial lead was placed in the right atrial appendage; the right ventricular (RV) lead was placed at the RV apex; the LV lead was implanted in the lateral or posterolateral vein. Echo optimization of atrioventricular and ventricular–ventricular timing was performed 2 weeks post-implantation using serial measurements of the aortic flow velocity envelopes.

Electrocardiogram definitions

All 12-lead ECGs were digitally read by a single observer who was blinded to the results of all investigations. The presence of LBBB or IVCD was defined by published guidelines.19 Further, LBBB was subdivided into those with an r wave in V1 ≥1 mm (r-V1) and/or a q wave in aVL ≥1 mm (q-aVL), residual conduction LBBB group (rLBBB), and those without either of these findings, cLBBB group (Figure 2).

Figure 2

Example tracings of complete left bundle branch block (cLBBB), left bundle branch block with residual left bundle conduction (rLBBB), and intraventricular conduction delay (IVCD). Note in complete left bundle branch block, the absence of an r wave in V1 or q wave in I, V6 and aVL. Left bundle branch block with residual left bundle conduction displays an r wave in V1 >1 mm, and a small q wave in aVL. The diagnosis of intraventricular conduction delay is made by the presence of a q wave in lead V6; note also the small r wave in V1.

Statistical analysis

Fisher's exact test was performed for comparison of proportions among groups. Welch's t-test was used for assessment of difference between means. A P value <0.05 was considered statistically significant. All statistical analysis was performed using Mathematical 8.01 (Wolfram Technologies, Champaign, IL, USA).


This study accords with the Declaration of Helsinki. The protocol received ethics approval from the University of Ottawa Heart Institute Research Ethics Board, and all patients signed informed consent. All potential conflicts of interest have been reported.



Fifty-one patients were recruited to PREDICT. One patient died before 3-month follow-up due to complications from a kidney infection (baseline ECG, LBBB with r in V1 >1 mm); another suffered recurrent lead dislodgement. Nine patients were excluded from this substudy due to the presence of RBBB, atrial fibrillation, or an in situ pacemaker. Thus, 40 patients were available for analysis and were classified into three groups: cLBBB (n = 12), rLBBB (n = 15), and IVCD (n = 13). Of patients with guideline LBBB, 11 were classified rLBBB by the presence of an r-V1 and 4 by the presence of q-aVL. The baseline characteristics of study patients are presented in Table 1. Left ventricular volumes [left ventricular end-systolic volume (LVESV) and left ventricular end-diastolic volume (LVEDV)] were significantly less in patients with cLBBB compared with rLBBB (P = 0.01). Scar burden (global, lateral, and septal) was similar between cLBBB and rLBBB. Intraventricular conduction delay patients had significantly greater septal scar than cLBBB patients (P = 0.01). Otherwise, patients were well matched for age, sex, LVEF, QRS duration, and baseline mechanical dyssynchrony between cLBBB and rLBBB.

View this table:
Table 1

Baseline variables stratified by the electrocardiogram group

VariablecLBBB (n= 12)rLBBB (n= 15)IVCD (n= 13)P value (cLBBB vs. rLBBB)P value (cLBBB vs. IVCD)
LVEF (%)24.66.320.
LVESV (mL)180.644.6258.189.9287.3152.10.010.03
LVEDV (mL)238.455.4318.990.5342.8148.90.010.03
QRS (ms)165.023.1163.722.4160.225.40.890.62
Mechanical dyssynchrony (°)a54.417.249.418.752.514.80.530.78
Global scar size (%)32.820.737.119.644.119.90.590.18
Lateral scar size (%)6.910.
Septal scar size (%)23.823.
Male7/12 (58%)13/15 (87%)12/13 (92%)0.190.07
Non-ischaemic aetiology7/12 (58%)5/15 (33%)7/13 (54%)0.261.0
Diabetes3/12 (25%)6/15 (40%)3/13 (23%)0.681.0
Hypertension8/12 (67%)7/15 (47%)4/13 (31%)0.440.12
  • LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; LVEDV, left ventricular end-diastolic volume; NYHA, New York Heart Association.

  • aMeasured in 11 patients.

Left ventricular remodelling

The mean absolute increase in LVEF, following CRT, for patients with cLBBB was 11.9 ± 11.9%. This compared with rLBBB, 3.8 ± 5.4% (P = 0.046) and IVCD, 2.5 ± 4.4% (P = 0.02 for cLBBB vs. IVCD). Mean percentage reduction in LVESV was 26.4 ± 39.2% in cLBBB, 14.3 ± 22.9% in rLBBB (P = 0.35), and 5.6 ± 17.3% in IVCD (P = 0.11 cLBBB vs. IVCD) (Table 2).

View this table:
Table 2

Change in variables (3 months post-cardiac resynchronization therapy compared with pre-implant) stratified by the electrocardiogram group

VariablecLBBB (n= 12)rLBBB (n= 15)NIVCD (n= 13)P value (cLBBB vs. rLBBB)P value (cLBBB vs. IVCD)
ΔLVEF (absolute %)11.911.
ΔLVESV (relative %)−26.439.2−14.322.9−5.617.30.350.11
ΔLVEDV (relative %)−16.435.7−10.422.3−3.415.10.630.26
ΔQRS (ms) native to native−8.011.0−
ΔQRS (ms) native to paced−14.020.9−
ΔMechanical dyssynchrony (°)−12.317.0−7.714.6−
  • LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; LVEDV, left ventricular end-diastolic volume; NYHA, New York Heart Association.

Clinical response

Clinical response was measured by change in NYHA class. There was a trend to a greater reduction in NYHA class in the cLBBB group following CRT, 0.67 ± 0.49, compared with rLBBB, 0.33 ± 0.49 (P = 0.09) (Table 2).

Electrocardiogram response to cardiac resynchronization therapy

Patients with cLBBB showed a strong trend towards a greater reduction in native QRS duration as compared with rLBBB: −8.0 ± 11.0 ms vs. −0.8 ± 8.2 ms (P = 0.07). Comparing cLBBB to IVCD (0.15 ± 8.1 ms) the difference was significant (P = 0.048). The change in QRS duration from native to CRT paced was not significantly different between the three groups (Table 2).

Change in mechanical dyssynchrony

There were reductions in mechanical dyssynchrony in all three groups following CRT (Table 2). Patients with cLBBB had the greatest reduction in mechanical dyssynchrony without reaching statistical significance.


The principal findings of this study are: patients with true cLBBB (as indicated by the absence of an r wave of ≥1 mm in V1 and/or a q wave ≥1 mm in aVL) had a greater increase in LVEF when compared with patients with ‘LBBB’ and residual LB conduction as suggested by the presence of these markers; further, there were trends to greater clinical response and reduced QRS duration with cLBBB.

QRS morphology is a strong predictor of response to CRT. It is well established that patients with RBBB or IVCD respond poorly,1518 while LBBB predicts a favourable response.3,18 In a recent examination of patients in the Multicenter Automatic Defibrillator Implantation Trial—Cardiac Resynchronization Therapy, patients with LBBB showed significantly greater degrees of LV remodelling compared with those with IVCD. Indeed, IVCD managed with CRT displayed a trend towards increased harm.18

However, recent data challenge current definitions of LBBB and IVCD. Padanilam et al.20 retrospectively analysed all cases of LBBB presenting to their electrophysiological laboratory where the right bundle branch was inadvertently damaged during the procedure. A significant proportion of their patients developed RBBB rather than complete heart block. A q wave in lead I or V5/V6 (which by current guidelines would define IVCD), and an r wave in V1 ≥1 mm (allowable by current guidelines for LBBB)19 predicted absence of complete heart block, and by implication residual LB conduction. One patient had a q wave in aVL ≥1 mm and likewise had residual LB conduction. However, r-V1 was the most sensitive ECG marker for this finding.

V1 is en face and proximate to the interventricular septum (Figure 1)—an r wave in V1 ≥1 mm in the context of ‘LBBB’ likely represents left-to-right septal activation as traditionally identified by the presence of q waves in V5/V6 and lead I. ‘LBBB’ with r-V1 and perhaps q-aVL may be better regarded as a form of IVCD. Compared with cLBBB, QRS widening in the context of IVCD is likely due to distal and widespread intramyocardial disease.23,24 We postulate that our patients may have had a lesser response to CRT for two reasons: a wide QRS in the context of residual LB conduction is a sign of more marked LV disease; residual LB conduction may reduce the septal-lateral electrical delay found in true cLBBB that is the ideal for resynchronization therapy.

Finally, we observed a greater reduction in native QRS duration (−8.0 ± 11.0 ms), following CRT, in patients with cLBBB compared with rLBBB. The effect of CRT on native QRS duration is controversial with some but not all studies suggesting a reduction in native QRS is usual.2528 Our findings suggest that QRS duration reduction in cLBBB patients is likely primarily due to a reduction in cardiac volumes.

Study limitations

The major limitation of this study is its small size that has resulted in large standard deviations compared with means. Therefore, small treatment effects may have been unobserved. Further, a small study size necessitated the use of a surrogate marker of CRT response—LV remodelling; we took this approach cognizant that is improvement in LV structure and function and not clinical response that is best correlated with reduced mortality, ventricular arrhythmia, and ICD shocks.811 However, for these reasons our work should be understood as hypothesis generating. A larger study looking at hard clinical endpoints is required before any clinical recommendations may be made.

An additional potential problem involves the mismatch of baseline LV volumes—smaller in cLBBB compared with rLBBB and IVCD. A previous publication found that baseline LV volume is not a predictor of CRT response and is therefore unlikely to have influenced our results.29


In guideline-defined LBBB the absence of an r wave in V1 ≥1 mm or a q wave in aVL ≥1 mm predicts a greater response to CRT. These results should be confirmed in a larger study.

Conflict of interest: M.S.G. is a speaker and received consulting honoraria from Medtronic, D.H.B. is a speaker and received honoraria from Medtronic, and research support from Medtronic and Boston Scientific. Other authors declare no conflict of interest.


This study was funded by a project grant from the J. P. Bickell Foundation (Toronto, Ontario) and was supported in part by a program grant from the Heart and Stroke Foundation of Ontario (HSFO, Ottawa, Ontario) (No. PRG6242).


View Abstract