Introduction: The mechanisms through which chest compressions (CC) generate blood flow remain controversial. The thoracic pump model is CC rate insensitive in the range 60 - 150 compressions per minute (cpm) but is sensitive to change in CC duty cycle with the prediction that increasing duty cycle would result in a modest, but linear, decrease in minute blood flow at a constant rate. Thus, the recently reported relationship between CC rate and outcomes may be due to changes in duty cycle instead of changes in rate. Methods: CPR was performed on nine domestic swine (∼30 Kg) using standard physiological monitoring. Flow was measured in the abdominal aorta, the inferior vena cava (IVC), the right renal artery and vein, the right common carotid and external jugular. Ventricular fibrillation (VF) was electrically induced. Mechanical CC were started after ten minutes of VF. CC were delivered at a rate of 50, 75, 100, 125, or 150 cpm and at a depth of 2” for a total of 54 min. CC rates were changed every 2 min and were randomized. The compression time was held constant at 350 ms. Results: CC generated blood flow showed a significant rate/duty cycle dependence and time dependence in all measured blood vessels. Early in the resuscitation, blood flows were optimized by rates above 100 cpm. As CC continued both the net blood flow and the optimal CC rate decreased as shown in Figure 1. Using previously collected data at 100 cpm and varying duty cycles, we found that minute blood flow did not decrease linearly with duty cycle, as predicted by the thoracic pump model. Conclusions: The optimal CC rate during CPR changes as a function of time suggesting that a fixed CC rate is not optimal. However, the data support the choice of ∼100 cpm as the “best” choice for a fixed CC rate. While some aspects of the observed rate effect may be explained by changes in CC duty cycle, these data do not support the idea that CC generated blood flow is independent of rate in the range 60-150 cpm.