Combination of current and new electrocardiographic-based criteria: a novel score for the discrimination of left bundle branch capture

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. Background Left bundle branch (LBB) area pacing (LBBAP) has been widely adopted for an increasing number of antibradycardia and CRT procedures worldwide. Two different types of capture have been described during LBBAP: left ventricular septal myocardial capture (LVSP) or direct LBB capture via non-selective (ns)-LBB or selective (s)-LBB pacing (LBBP). Nevertheless, electrocardiographic diagnosis of direct LBB capture remains one of the challenges of modern conduction system pacing. Purpose We hypothesized that the combination of several ECG-based criteria might discriminate better LBBP from LVSP, than each criterion separately. Methods Single-center study involving all consecutive patients who received LBBAP. LBBP was defined according to QRS morphology transition criteria during decremental pacing. Multivariate logistic regression analysis was performed to develop a predictive score for LBBP (The LBBP score), by entering three ECG-based criteria: the previously reported and widely adopted R wave peak time in lead V6 (V6-RWPT) and V6-V1 interpeak interval, and the novel aVL-RWPT, that had been tested before in our sample by the performance of ROC curve analysis. The predictive performance of the regression model was evaluated through the ROC curve and compared to those of the isolated criteria with the De Long test. The regression equation derived from the multivariate model was obtained. To simplify its performance in clinical practice, a heuristic method was used to develop the score: each criterion was weighted according to the result of the Wald test. The construction of the score is depicted in Figure 2. Results A total of 188 patients with intended LBBAP were screened. Successful LBBAP was achieved in 174 (92.5%). 71 patients with confirmed LBB capture by the QRS morphology transition criteria were analysed to develop the predictive score. The optimal cut-off values of V6-RWPT, V6-V1 interpeak interval and aVL-RWPT for the discrimination of LBBP were <83 ms, ≥33 ms and <79 ms, respectively. The multivariate logistic regression model showed that the three analysed ECG-based criteria were independent and not redundant predictors of LBBP. The ROC analysis for the multivariate regression model showed an AUC of 0.980 to differentiate LBB and LVS captures. The probability of LBBP can be calculated by using the calculation formula shown in Figure 1. The ROC curve for the differential diagnosis of LBBP and LVSP with the score showed an AUC of 0.976, with an optimal cut-off value of 3 points (Sensitivity 89.2%, Specificity 100%) for the differentiation of LBB capture. Conclusions The combination of V6-RWPT, V6-V1 interpeak interval and aVL-RWPT into a new score exhibited a good power to discriminate LBBP from LVSP. None of the reported cut-off values of current ECG-based criteria have shown such high relationship between both sensitivity and specificity, so the LBBP score might be an efficient and precise tool to implement in clinical practice.