In patients with sinus node dysfunction (SND) with or without associated paroxysmal atrial fibrillation (AF), the effectiveness of atrial pacing in reducing the incidence of AF is not definitive. In addition, despite several studies involving large populations of implanted patients, little attention has been paid to the electrophysiological (EP) atrial substrate and the effect of permanent atrial pacing. The aim of this study is to correlate EP data and the risk of AF after DDD device implantation. We reviewed EP data of 38 consecutive patients with SND, mean age 70 +/- 8 years, who were investigated free of antiarrhythmic treatment, for the evaluation of the atrial substrate. We also considered as control group 25 subjects, mean age 63 +/- 14 years, referred to our EP laboratory for unexplained syncope or various atrioventricular disturbances. Following pharmacological washout and at a drive cycle length of 600 ms, effective and functional refractory periods (ERP, FRP), S1-A1 and S2-A2 latency, A1 and A2 conduction duration, and latent vulnerability index (ERP/A2) were measured. AF induction was tested with up to three extrastimuli at paced cycle lengths of 600 and 400 ms in 20 patients. Induction of sustained AF (> 30 seconds) was considered as the endpoint. P wave duration on the surface ECG in lead II/V1 was also measured. DDD pacing mode was chosen in all patients with the minimal atrial rate programmed between 60 and 75 beats/min (mean 64 +/- 4 beats/min). After implantation, the patients were followed-up for 29 +/- 17 months and clinically documented occurrence of AF was determined. When comparing patients with SND and subjects of the control group, we did not find any significant statistical differences in terms of ERP (237 +/- 33 vs 250 +/- 29 ms), FRP (276 +/- 30 vs 280 +/- 32 ms) and S1-A1 (39 +/- 16 vs 33 +/- 11 ms) and S2-A2 latency (69 +/- 24 vs 63 +/- 25 ms). In contrast, we observed significant differences regarding A1 (55 +/- 19 vs 39 +/- 13 ms; P < 0.001), A2 (95 +/- 34 vs 57 +/- 18 ms; P < 0.001) and P wave duration (104 +/- 18 vs 94 +/- 15 ms; P < 0.05), and ERP/A2 (2.8 +/- 1.2 vs 4.8 +/- 1.6; P < 0.001). When comparing patients with (n = 11) or without (n = 27) postpacing AF occurrence, we did not find any difference with reference to ERP, FRP, S1-A1, S2-A2, A1 duration, or follow-up duration. In patients with postpacing AF occurrence, A2 was longer (116 +/- 41 vs 87 +/- 27 ms; P < 0.01), ERP/A2 lower (2.1 +/- 0.4 vs 3.1 +/- 1.4; P < 0.05), P wave more prolonged (116 +/- 22 vs 99 +/- 14 ms; P < 0.01), and preexisting AF history predominant (6/11 vs 5/27 patients; P < 0.05). No difference was observed between patients with (n = 8) and without (n = 12) AF induction during the EP study. In patients with SND, the atrial refractoriness appears normal and the most important abnormality concerns conduction slowing disturbances. Persistence of AF despite pacing stresses the importance of mechanisms responsible for AF not entirely brady-dependent. In this setting, more prolonged atrial conduction disturbances, responsible for a low vulnerability index, and a preexisting history of AF enable us to identify a high risk patient group for AF in the follow-up.