BACKGROUND Selective atrioventricular nodal (AVN) vagal stimulation (AVN-VS) has emerged as a novel strategy for ventricular rate (VR) control in atrial fibrillation (AF). Although AVN-VS preserves the physiological ventricular activation sequence, the resulting rate is slow but irregular. In contrast, AVN ablation with pacemaker implantation produces retrograde activation (starting at the apex), with regular ventricular rhythm. We tested the hypothesis that, at comparable levels of VR slowing, AVN-VS provides hemodynamic benefits similar to those of ablation with pacemaker implantation. RESULTS AVN-VS was delivered to the epicardial fat pad that projects parasympathetic nerve fibers to the AVN in 12 dogs during AF. A computer-controlled algorithm adjusted AVN-VS beat by beat to achieve a mean ventricular RR interval of 75%, 100%, 125%, or 150% of spontaneous sinus cycle length. The AVN was then ablated, and the right ventricular (RV) apex was paced either irregularly (i-RVP) using the RR intervals collected during AVN-VS or regularly (r-RVP) at the corresponding mean RR. The results indicated that all 3 strategies improved hemodynamics compared with AF. However, AVN-VS resulted in significantly better responses than either r-RVP or i-RVP. i-RVP resulted in worse hemodynamic responses than r-RVP. The differences among these modes became less significant when mean VR was slowed to 150% of sinus cycle length. CONCLUSIONS AVN-VS can produce graded slowing of the VR during AF without destroying the AVN. It was hemodynamically superior to AVN ablation with either r-RVP or i-RVP, indicating that the benefits of preserving the physiological antegrade ventricular activation sequence outweigh the detrimental effect of irregularity.