Aim: The recurrence of atrial fibrillation after cardioversion was a general problem in medical practice. Methods: We wanted to test the hypothesis that the recurrence of atrial fibrillation could be predicted by measurements of P wave triggered P wave signal averaged ECG and chemoreflex sensitivity (CHRS) in 118 consecutive patients one day after successful electrical cardioversion. We measured the filtered P wave duration (FPD) and the root mean square voltage of the last 20 ins of the P wave (RMS 20) with P wave triggered P wave signal averaged ECG. The CHRS was calculated of a ratio between the difference of RR intervals in ECG and venous pO(2) before and after 5-min oxygen inhalation. The pathologic CHRS was predefined as a CHRS below 3.0 ms/mm Hg. Results: A recurrence of atrial fibrillation was observed in 57 patients (48%) during a 9.3-month follow-up. The left atrial size (41.9 +/- 4.0 vs. 39.3 +/- 3.1 mm, P < 0.0003) was larger, the FPD (139.6 +/- 16.0 vs. 118.2 +/- 14.1 ms, P < 0.0001) was longer, the RMS 20 (2.57 +/- 0.77 vs. 3.90 +/- 0.99 mu V, P < 0.0001) and the CHRS (2.66 +/- 1.18 vs. 4.01 +/- 1.66 ms/min Hg, P < 0.0001) were lower in patients with recurrence of atrial fibrillation. A cut-off point was defined with FPD >= 126 ms and RMS 20 <= 3.1 mu V and could predict atrial fibrillation with a specificity of 77%, a sensitivity of 72%, a positive value of 75%, a negative predictive value of 75% and an accuracy of 75%. Patients with FPD > 126 ms and RMS 20 <= 3.1 mu V had nearly a 7-fold increase risk for recurrence of atrial fibrillation. Conclusion: The recurrence of atrial fibrillation after cardioversion could be detected by P wave signal averaged ECG and an analysis of CHRS. These methods seem to be appropriate to show a delayed atrial conduction and a neurovegetative imbalance which might be possible trigger mechanisms of reinitiating of atrial fibrillation. (c) 2005 Published by Elsevier Ireland Ltd.