We tested whether biventricular resynchronization explains contractile function changes with univentricular and biventricular pacing in heart failure patients with varying magnitudes of baseline biventricular asynchrony. Thirty patients (New York Hospital Association class > or = III, QRS duration > or =120 ms) were tested. Contractile function was measured by left ventricular maximum first derivative of pressure over time (dP/dtmax). Biventricular mechanical asynchrony was quantified by the normalized pressure-pressure (NPP) loop area formed by the cross-plot of right and left intraventricular pressure curves from each cardiac cycle. Any ventricular pacing increased dP/dtmax if it decreased baseline NPP loop area and almost always worsened dP/dtmax and asynchrony when baseline NPP loop area <0.3. The quantitative relationship between dP/dtmax and NPP loop area change depended on ventricular pacing site and timing relative to intrinsic activation. For similar NPP loop decreases, dP/dtmax increased 16% more with left and biventricular pacing compared with right ventricular pacing. In conclusion, right, left, or biventricular pacing can improve contractile function only in patients having sufficient baseline biventricular asynchrony. However, biventricular resynchronization is only one of the improvement mechanisms.