Prediction of maintenance of sinus rhythm after cardioversion of atrial fibrillation by analysis of serial signal-averaged P waves.

After cardioversion from atrial fibrillation (AF) many patients develop early recurrence of the arrhythmia. While these patients may be appropriate for immediate prophylaxis against AF recurrence their identification at the time of cardioversion is not possible. Since the signal-averaged P wave (SAPW) is abnormal in individuals with atrial arrhythmia, we assessed its utility for predicting early AF recurrence after cardioversion. Seventy-five cardioversions in 31 patients were evaluated. The mean age was 59 (range 28-79) years; 26 were male. Fifty-eight cardioversions were internal using low energy biphasic DC shocks delivered via electrodes placed in the right atrial appendage and coronary sinus. P wave specific signal averaging was performed at 3 and 24 hours after each cardioversion to estimate filtered P wave duration and energy from 20, 40, and 60 to 150 Hz. Follow-up was by regular clinic visits and transtelephonic ECG monitoring. Early recurrence of AF (prospectively defined as sinus rhythm duration < 1 week) occurred after 30 cardioversions. No differences were found in any P wave variable measured at 3 hours between these cardioversions and those that resulted in a longer duration of sinus rhythm. Paired 3- and 24-hour signal-averaged data were available in 47 cardioversions. There were significant falls in P wave energy from 3 to 24 hours after 31 cardioversions that resulted in sinus rhythm for > 1 week, (P40: 3 hours 11.2 [+/- 1.5] micro V2.s, 24 hours 8.6 [+/- 1.2] micro V2.s, P < 0.001), but not following the 16 after which AF returned within 1 week (P40: 3 hours 9.0 [+/- 1.2] micro V2.s, 24 hours 8.5 [+/- 1.2 micro V2.s, P = NS). A fall in P40 of > 25% had a positive predictive accuracy for maintenance of sinus rhythm of 87%; negative predictive accuracy was only 37%. Similar falls in P wave energy occurred after cardioversions that resulted in longer term (> 4 weeks) sinus rhythm, but not in those that did not. However, the predictive accuracy of a fall in P40 was less (positive predictive accuracy 38%, negative predictive accuracy 62%). Patients with relapsing permanent AF who remain in sinus rhythm for at least 1 week after cardioversion show a fall in P wave energy within the first 24 hours. However, in these patients the technique does not predict recurrent AF within 1 week nor sinus rhythm > 4 weeks. These observations suggest persistent disordered atrial activation as a mechanism for early recurrence of AF after cardioversion.