Autothreshold algorithm feasibility and safety in left bundle branch pacing

Aims: Autothreshold algorithms enable remote monitoring of patients with conventional pacing, but there is limited information on their performance in left bundle branch pacing (LBBP). Our objective was to analyse the behaviour of the autothreshold algorithm in LBBP and compare it with conventional pacing and manual thresholds during initial device programming (acute phase), after 1-7 days (subacute), and 1-3 months later (chronic).

Methods and results: A prospective, non-randomized, single-centre comparative study was conducted. Consecutive patients with indication for cardiac pacing were enrolled. Implants were performed in the left bundle branch area or the right ventricle endocardium at the discretion of the operator. Left bundle branch pacing was determined according to published criteria. Autothreshold algorithm was activated in both groups whenever allowed by the device. Seventy-five patients were included, with 50 undergoing LBBP and 25 receiving conventional pacing. Activation of the autothreshold algorithm was more feasible in later phases, showing a favourable trend towards bipolar pacing. Failures in algorithm activation were primarily due to insufficient safety margins (82.8% in LBBP and 90% in conventional pacing). The remainder was attributed to atrial tachyarrhythmias (10.3% and 10%, respectively) and electrical noise (the remaining 6.9% in the LBBP group). In the LBBP group, there were not statistically significant differences between manual and automatic thresholds, and both remained stable during follow-up (mean increase of 0.50 V).

Conclusion: The autothreshold algorithm is feasible in LBBP, with a favourable trend towards bipolar pacing. Automatic thresholds are similar to manual in patients with LBBP, and they remain stable during follow-up.