Spatial and temporal stability of the dominant frequency of activation in human atrial fibrillation.

PURPOSE

Intraoperative activation sequence mapping during atrial fibrillation (AF) is difficult because of the complexity of the data. The data analysis is time consuming, and picking activation times can be ambiguous. The purpose of this study was to determine whether mapping in the frequency domain during AF can be used to rapidly locate the region and assess the stability of the dominant frequency.

METHODS

In 33 patients, epicardial bipolar electrograms were intraoperatively recorded from 250 sites during AF. For each electrogram, a power spectrum was calculated using a fast Fourier transform. The peak frequency below 11 Hz was determined from the power spectrum for each electrogram.

RESULTS

Isofrequency mapping demonstrated that 91% of the patients exhibited a distinct region of maximum (dominant) frequency at least once during the recording period. Nine percent had no distinct region of dominant frequency. A distinct region of stable dominant frequency was located in the left atrium 30% of the time and the right atrium 12% of the time for the entire recording period. The location of dominant frequency changed during the recording period in 48% of the patients. The dominant frequency was highest in patients with chronic AF (8.29 +/- 1.0 Hz) compared to those with paroxysmal AF (PAF) (6.54 +/- 0.62 Hz, P = .001). The dispersion of frequency was higher in the patients with persistent AF (1.03 +/- 0.4 Hz) than in those with PAF (0.6 +/- 0.3 Hz, P < .001).

CONCLUSION

Frequency mapping rapidly and accurately identifies the region of dominant activation frequency. The frequency is faster and more variable in persistent AF than in PAF. The location of the dominant frequency was unstable, changing during the recording period, in half the patients. The location of the dominant frequency was independent of the type of AF.