Background: The effects of atrial pacing mode on atrial and ventricular function in patients with atrial fibrillation (AF) and bradycardia have not been evaluated. We evaluated atrial and ventricular function during randomization to support pacing (SP), high right atrial pacing (HRA), and dual site right atrial pacing (DAP). Methods: Seventy-nine patients (66 ± 12 yr, 46 male) with standard pacing indications and symptomatic AF were randomized to each of three pacing modes (DAP, HRA, SP) for 6 months in a crossover design. Echocardiographic studies were performed at enrollment and the end of each mode. Paired comparisons of atrial and ventricular function parameters were performed between each pacing mode and baseline. Results: HRA pacing in DDDR mode resulted in increased left ventricular (LV) end systolic volume (78 ± 42 vs. 60 ± 31 ml, p = 0.001) and reduced LV ejection fraction (44 ± 14 vs. 50 ± 11%, p = 0.007) compared to baseline. These parameters did not change during DAP. DAP resulted in increased peak A wave velocity (75 ± 19 vs. 63 ± 23 cm/s, p = 0.003) and atrial filling fraction compared to baseline (0.47 ± 0.15 vs. 0.38 ± 0.13, p = 0.005). Atrial and ventricular function were similar between control and SP. Conclusion: DAP, but not HRA or SP, improved left atrial (LA) function in patients with AF and bradycardia. HRA pacing in DDDR mode resulted in LA dilatation and deterioration of LV function which was not observed with DAP. © 2005 Springer Science + Business Media, Inc.