Permanent biatrial and/or multisite atrial pacing may prevent atrial fibrillation (AF), but the effects on atrial electrophysiology remain incompletely understood. Acute biatrial pacing was studied in 20 patients with and 28 without (controls) a history of atrial fibrillation and/or flutter. Twelve-lead electrocardiograms were recorded during pacing from the high right atrium (RA), from the distal coronary sinus (LA), and biatrial pacing. P wave duration was measured in each lead and the difference between maximum and minimum P duration was termed P wave dispersion. Effective refractory periods (ERPs) were measured during each pacing mode. The dispersion of P wave duration was 35 +/- 14 ms in controls and 40 +/- 29 ms in AF patients (P = 0.17). Compared to RA pacing, LA pacing shortened P duration in controls (127 +/- 18 to 107 +/- 16 ms, P < 0.05) and biatrial pacing markedly shortened P duration in controls (127 +/- 18 to 93 +/- 14 ms, P < 0.05) and AF patients (114 +/- 43 to 97 +/- 21 ms, P < 0.05). P wave dispersion was unaffected. In controls, the LA ERP was longer than the RA ERP. This phenomenon was not present in AF patients, whose LA ERP was shorter than that of controls. Biatrial pacing had no effect on atrial ERPs or the dispersion of atrial refractoriness. In conclusion, acute biatrial pacing does not affect atrial repolarization but it does cause a marked shortening of global biatrial depolarization. Distal coronary sinus pacing produces a shorter P wave than RA pacing. There is substantial dispersion in the surface P wave of the electrocardiogram, the significance of which awaits further study.