Bimodal distribution of QRS complex duration in patients with dilated cardiomyopathy undergoing an ICD implant procedure

Abstract

Abstract Background/Introduction In patients with heart failure with reduced ejection fraction (HFrEF), QRS duration is of the uttermost importance. Substantial proportion of patients with HFrEF and sufficiently prolonged QRS complex, especially those with left bundle branch block (LBBB), may significantly benefit from cardiac resynchronization therapy. Purpose The aim of the study was to analyse the distribution of the QRS durations in patients with HFrEF undergoing ICD implant procedure to better understand the phenomenon of QRS prolongation. Methods All patients undergoing the implant procedure in our centre between January 2017 and December 2021 for HFrEF caused by coronary artery disease (CAD) or dilated cardiomyopathy (DCM) were analysed for the inclusion into the study. The exclusion criteria were (1) coincidence of CAD with DCM, (2) upgrade from any other cardiac rhythm management device, (3) unavailability of the ECG recording or ECG recording with insufficient quality before the implant procedure, (4) more than one ventricular premature beat in the recording. The study used automatically measured QRS duration by Mortara EKG machines (299 recordings with ELI350, 98 with ELI380, and 10 with ELI250). Statistical analysis was performed in R software [1] with the function locmodes for the visualisation of probability density and excess mass test [2] for assessing statistical significance, both from the multimode package. Results The study included 407 patients. In 113 patients with DCM and 177 patients with CAD, the implantation was based on primary prevention (PP) indications, and 24 and 93 patients with DCM and CAD were implanted for secondary prevention (SP) indications. The distribution of QRS durations in the CAD-PP, as well as CAD-SP was unimodal (p=ns). In the DCM-PP group, the distribution was bimodal (p=0.012), in the DCM-SP group, the trend was non-significantly bimodal (p=0.076). Both DCM groups taken together had clearly bimodal distribution (p=0.002 – probability density curve according to locmodes function in the figure). The values between 139 and 142 ms were not present in this group. Below this border zone, two other patients fulfilled the criteria of LBBB, both with the duration close to the border zone (137 and 138 ms). Above this zone, 8 patients had right bundle branch block, the rest had LBBB. Conclusions Bimodal distribution of QRS duration values in patients with DCM supports the idea of sudden change of left bundle conduction rather than slow progression of local changes. In CAD, scars and other local impairments in conduction may hide these changes and no clear border zone can be set. With the well-known differences in the measurements of main ECGmachine manufacturers, criteria for bundle branch block in DCM could be set with high accuracy. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Charles University Research Program “Cooperatio – Cardiovascular Science” and the Ministry of Health of the Czech Republic Conceptual Development of Research Organization