Clinical and electrocardiographic predictors of recurrent atrial fibrillation.

Patients with frequent episodes of paroxysmal atrial fibrillation (AF) are prone to develop permanent AF and have an increased thromboembolic risk. We have previously shown that P wave dispersion (P dispersion), defined as the difference between the maximum and the minimum P wave duration, and maximum P wave duration (P maximum) can distinguish patients with paroxysmal lone AF. The ability of those ECG markers and of other clinical and ECG variables to detect patients at risk for recurrent AF was tested in 88 patients, aged 64 +/- 12 years. All patients had a history of symptomatic episodes of AF during the last 2 years and had not previously received any antiarrhythmic prophylaxis. P maximum and P dispersion were calculated from a 12-lead surface ECG recorded in all patients during sinus rhythm. A computerized ECG system was used and P maximum and P dispersion were calculated on screen from the averaged complexes of all 12 leads. Age (P = 0.01), history of organic heart disease (P = 0.03), P maximum (P < 0.001), minimum P wave duration (P = 0.05), and P dispersion (P < 0.001) were found to be significant univariate predictors of recurrent AF, whereas only P maximum (P < 0.001) and age (P = 0.037) remained significant independent predictors of frequent AF paroxysms in the multivariate analysis. It is concluded that advanced age and prolonged P wave duration may be used as predictors of frequently relapsing AF. Therefore, simple AF predictors exist that could possibly distinguish the patients in whom prophylaxis with antiarrhythmic medicines should be instituted.