Aims: The study set out to explore whether an index of atrial electrical electrophysiology can be used to predict atrial fibrillation (AF) relapse, and if the predictive properties differ as a result of arrhythmia duration.
Methods and results: The study comprised 175 consecutive patients with persistent AF (median duration 94 days, range 2 to 1044) referred for cardioversion. Twenty-nine patients had arrhythmia duration under 30 days (median 5 days, range 2-26). Atrial fibrillatory rate (AFR) was estimated using a frequency power spectrum analysis of QRST-cancelled ECG. At 1-month follow-up, 56% of the patients had relapsed to AF. The pre-cardioversion mean AFR of those patients was 399+/-52 fibrillations per minute (fpm) compared with 363+/-63 fpm among patients maintaining SR (P<0.0001). In patients with short AF duration, the difference was even more pronounced (424+/-52 vs. 345+/-65 fpm, P<0.01). In this group, a finding of an AFR above the mean value of the study population predicted AF relapse with high accuracy.
Conclusion: In patients undergoing cardioversion of persistent AF, AF relapse is predicted by a higher AFR. A stronger association is seen in patients with short arrhythmia duration, reflecting either rapid remodelling or pre-existing changes in those who relapse to AF.