The purpose of this randomized, prospective trial was to determine if Bachmann's bundle pacing reduces the incidence of AF after CABG. The study included 161 patients with no history of AF who were randomized to three groups. Group 1 included 50 patients as controls. Group 2 included 60 patients who had an epicardial wire placed at the lateral wall of the right atrium. In the 51 patients of group 3, the wire was placed at the Bachmann's bundle. In groups 2 and 3, atrial pacing (AAI 96 beats/min) was initiated immediately after CABG and continued for 5 days. The study endpoint was AF lasting > or = 1 minute. Baseline clinical parameters were similar in all three groups. The incidence of AF was not reduced by pacing (group 1: 42%; group 2:48%; group 3:37%; P = NS). The paced P wave duration was increased in group 2 (129 +/- 14 ms vs group 3: 96 +/- 21 ms; P < 0.05). Paced P wave duration was a risk factor for postoperative AF (odds ratio 1.015; 95% CI 1.0021-1.028; P < 0.05). Analysis comparing the pacing groups revealed a reduction in AF during Bachmann's bundle pacing (50 vs 29%; P < 0.01). Pacing thresholds were significantly better at Bachmann's bundle compared to group 2. In conclusion, an anatomically guided pacing at the Bachmann's bundle does not reduce the overall incidence of postoperative AF compared to controls. However, the Bachmann's bundle offers favorable capabilities for postoperative a trial pacing, and thus it is a preferable site for electrode placement if postoperative atrial pacing is required.