Atrial Conduction Velocity Correlates with Frequency Content of Bipolar Signal

Background: Anisotropy in conduction velocity (CV) is a key substrate abnormality influencing atrial arrhythmias. In skeletal muscle fibers, CV and frequency content of the surface electromyogram signal are directly related. We hypothesized that in human atria the frequency content of the bipolar signal, recorded on the endocardial surface, is directly related to the local CV. Methods: In 15 patients submitted to ablation of supraventricular arrhythmias, incremental pacing was performed through an octapolar catheter inserted into the coronary sinus (CS), alternatively from both extremities in two different sequences: CS bipole 1–2 as the pacing site and CS bipole 7–8 as the detection site in the first, and vice versa in the second. The pacing cycle length (PCL) was stepwise decreased from 600 ms to 500 ms, 400 ms, 300 ms, until 250 ms. Estimation of the CV was performed as the ratio between the distance traveled by the propagating pulse and the propagation time. The frequency distribution of the signal energy was estimated using the fast Fourier transform, and the characteristic frequency (CF) was estimated as the barycenter of the frequency spectrum. Results: A total of 2,496 bipolar signals were analyzed; CV and CF were estimated and compared. The single patient and group data analysis showed a significant direct correlation between CV and CF of the local bipolar signal. Conclusions: Comparing the degree of spectral compression among signals registered in different points of the endocardial cardiac surface in response to decreasing PCL enables to map local differences in CV, a useful arrhythmogenic substrate index.