Partial versus complete left bundle branch block, does it impact the outcomes with cardiac resynchronization therapy?

The prevalence of a significant percentage of “non-responders” to cardiac resynchronization therapy (CRT), i.e. patients who demonstrate no improvement in heart failure outcomes despite institution of CRT with biventricular pacing [1,2], indicates that the method of selecting patients for CRT is not yet fully adequate. Just as important as identifying who will benefit from CRT is knowing in whom it can safely be deferred, especially with the development of conduction system pacing, an alternative method to restore physiologic ventricular synchrony. A subgroup analysis of the PREDICT study examining forty patients showed a greater reduction in LV EF and end systolic volume as well as NYHA class in patients with complete left bundle branch block (LBBB) in comparison to patients with residual conduction over the left bundle (rLBBB) [3].

To that end, in the context of the afore-mentioned PREDICT substudy, the authors of the present study performed a retrospective observational cohort study to evaluate how patients at their center with rLBBB fared after CRT in comparison with patients who had a true LBBB. Grebmer and colleagues included in their study all 690 consecutive patients who underwent CRT implant over a seven year period at a center in Munich, Germany. Of the cohort, 52% had a true LBBB; 11% had a LBBB with residual conduction, 19% had a pacing-induced “LBBB”, and the remaining patients had an IVCD with the exception of the <1% who had a right bundle branch block. Complete LBBB and residual conduction over the left bundle were distinguished on 12 lead ECG using the same criteria as in the PREDICT substudy. Between patients who had a true LBBB or LBBB with residual conduction there was no significant different in mortality over the follow up period. There was also no significant difference between the two groups in improvement in left ventricular ejection fraction at one year post-implant. Both the LBBB and rLBBB groups were very similar with respect to baseline characteristics (age, pre-implant QRS duration, baseline LV EF, and the prevalence of diabetes, renal failure, and male gender) and there is no plausible reason, especially outside the context of a prospective study, for the two groups to have received differing clinical care on the basis of their conduction abnormalities. The post implant QRS between the two groups was similar as well (mean QRS 137.3 ms vs 137.1 ms). Consistent with the results of prior studies, the group of patients in the present cohort who had an intraventricular conduction delay (IVCD) at baseline did not demonstrate an improvement in LV EF after CRT, and in fact showed increased mortality in comparison to patients with LBBB, acknowledging that the IVCD group had an increased prevalence of diabetes, renal failure, and male gender-all predictors of poorer outcomes.

Grebner and colleagues note the absence of consistent documentation regarding lead position as a particular limitation of the study. However, they note that their standard practice to target basal or mid lateral branches of the coronary sinus for lead delivery and the clinical (rather than academic) motivation in achieving adequate lead position lends to the “real-world” nature of the cohort. Some other limitations of the study deserve mention, apart from those acknowledged by the authors. Survival data were reportedly missing for nearly a quarter of patients, but the handling of missing data was not commented on by the authors. Were survival data disproportionately missing for one group over others and how was the missing data accounted for in the analysis? Unrelatedly, while the inclusion of a group of patients with IVCDs was necessary to serve as a control group, one needs to bear in mind that this analysis took place during a time when patients with IVCDs were known to have reduced response to CRT, and it was likely a rather selected population of patients with IVCDs.

Such biases and limitations are not uncommon in retrospective observational studies. Nevertheless, the study design was sufficient as a next step in evaluating whether the PREDICT subgroup analysis finding of reduced benefit from CRT among those with LBBB but residual conduction over the left bundle warranted more intense study with a more rigorous design. In the face of the present study's results, the result of that subgroup analysis may be relegated to the heap of other subgroup analysis findings which have not borne out in subsequent study.