Background— We sought to compare left ventricular (LV _epi ) and biventricular epicardial pacing (BIV _epi ) with LV (LV _endo ) and BIV endocardial pacing (BIV _endo ) in patients with chronic heart failure with an emphasis on the underlying electrophysiological mechanisms and hemodynamic effects. Methods and Results— Ten patients with chronically implanted cardiac resynchronization devices underwent temporary LV _endo and BIV _endo pacing with an LV endocardial roving catheter. A pressure wire and noncontact mapping array were placed to the LV cavity to measure LVdP/dt _max and perform electroanatomical mapping. At the optimal endocardial position, the acute hemodynamic response (AHR) was superior to epicardial stimulation, the AHR to BIV _endo pacing and LV _endo pacing being comparable (21±15% versus 22±17%; P =NS). During intrinsic conduction, QRS duration was 185±30 ms, endocardial LV total activation time 92±27 ms, and trans-septal activation time 60±21 ms. With LV _endo pacing, QRS duration (187±29 ms; P =NS) and endocardial LV total activation time (91±23 ms; P =NS) were comparable with intrinsic conduction. There was no significant difference in endocardial LV total activation time between LV _endo and BIV _endo pacing (91±23 versus 85±15 ms; P =NS). Assessment of isochronal maps identified slow trans-septal conduction with both LV _endo and BIV _endo pacing resulting in activation of almost the entire LV endocardium prior to septal breakout, thereby limiting any possible fusion with either pacing mode. Conclusions— The equivalent AHR to LV _endo and BIV _endo pacing may be explained by prolonged trans-septal conduction limiting fusion of electrical wavefronts. The optimal AHR was associated with predominantly LV pre-excitation and depolarization. Our results suggest that LV pacing alone may offer a viable endocardial stimulation strategy to achieve cardiac resynchronization.