Algorithm for ventricular capture verification based on the mechanical evoked response

Abstract

Automatic pacemaker capture verification is important for maintaining safety and low energy consumption in pacemaker patients. A new algorithm was developed, based on impedance measurement between pacing electrode poles, which reflects the distribution of the conducting medium between the poles and changes with effective contraction. Data acquired during pacemaker implant in 17 subjects were analysed, with intracardiac impedance recorded while pacing was performed in the ventricle at varying energies, resulting in multiple-captured and non-captured beats. The impedance signals of all captured/non-captured beats were analysed using three different algorithms, based on the morphology of the impedance signal. The algorithm decision for each beat was compared with an actual capture or non-capture, as determined from the simultaneous recording of surface ECG. Two of the three algorithms (Z1 and Zn) were based on impedance values, and one (Z'n) was based on the first derivative of the impedance. Z1 was based on a single sample, whereas Z'n and Z'n were based on several samples in each beat. The total accuracy for each was Z1: 43%, Zn: 87%, Z'n: 92%. It was concluded that impedance-based capture verification is feasible, that a multiple rather than single sample approach for signal classification is both feasible and superior, and that first derivative analysis with multiple samples (Z'n) provides the best results.