We evaluated the acute impact of different cardiac pacing sites on two-dimensional speckle-tracking echocardiography (STE) derived left ventricular (LV) rotation and twist in healthy dogs.Twelve dogs were used in this study. The steerable pacing electrodes were positioned into right heart through the superior or inferior vena cava, into LV through aorta across the aortic valve. The steerable pacing electrodes were positioned individually in the right atrium (RA), right ventricular apex (RVA), RV outflow tract (RVOT), His bundle (HB), LV apex (LVA) and LV high septum (LVS), individual pacing mode was applied at 10 minutes interval for at least 5 minutes from each position under fluoroscopy and ultrasound guidance and at stabilized hemodynamic conditions. LV short-axis images at the apical and basal levels were obtained during sinus rhythm and pacing. Offline STE analysis was performed. Rotation, twist, time to peak rotation (TPR), time to peak twist (TPT), and apical-basal rotation delay (rotational synchronization index, RSI) values were compared at various conditions. LV pressure was monitored simultaneously.Anesthetic death occurred in 1 dog, and another dog was excluded because of bad imaging quality. Data from 10 dogs were analyzed. RVA, RVOT, HB, LVA, LVS, RARV (RA+RVA) pacing resulted in significantly reduced apical and basal rotation and twist, significantly prolonged apical TPR, TPT and RSI compared to pre-pacing and RA pacing (all P<0.05). The apical and basal rotation and twist values were significantly higher during HB pacing than during pacing at ventricular sites (all P<0.05, except basal rotation at RVA pacing). The apical TPR during HB pacing was significantly shorter than during RVOT and RVA pacing (both P<0.05). The LV end systolic pressure (LVESP) was significantly lower during ventricular pacing than during pre-pacing and RA pacing.Our results show that RA and HB pacing results in less acute reduction on LV twist, rotation and LVESP compared to ventricular pacing.