This study compares LV performance during high right ventricular septal (RVS) and apical (RVA) pacing in patients with LV dysfunction who underwent His-bundle ablation for chronic AF. We inserted a passive fixation pacing electrode into the RVA and an active fixation electrode in the RVS. A dual chamber, rate responsive pulse generator stimulated the RVA through the ventricular port and the RVS via the atrial port. Patients were randomized to initial RVA (VVIR) or RVS (AAIR) pacing for 2 months. The pacing site was reversed during the next 2 months. At the 2 and 4 month follow-up visit, each patient underwent a transthoracic echocardiographical study and a rest/exercise first pass radionuclide ventriculogram. We studied nine men and three women (mean age of 68 +/- 7 years) with congestive heart failure functional Class (NYHA Classification): I (3 patients), II (7 patients), and III (2 patients). The QRS duration was shorter during RVS stimulation (158 +/- 10 vs 170 +/- 11 ms, P < 0.001). Chronic capture threshold and lead impedance did not significantly differ. LV fractional shortening improved during RVS pacing (0.31 +/- 0.05 vs 0.26 +/- 0.07, P < 0.01). RVS activation increased the resting first pass LV ejection fraction (0.51 +/- 0.14 vs 0.43 +/- 0.10, P < 0.01). No significant difference was observed during RVS and RVA pacing in the exercise time (5.6 +/- 3.2 vs 5.4 +/- 3.1, P = 0.6) or the exercise first pass LV ejection fraction (0.58 +/- 0.15 vs 0.55 +/- 0.16, P = 0.2). The relative changes in QRS duration and LV ejection fraction at both pacing sites showed a significant correlation (P < 0.01). We conclude that RVS pacing produces shorter QRS duration and better chronic LV function than RVA pacing in patients with mild to moderate LV dysfunction and chronic AF after His-bundle ablation.