BACKGROUND The number of chest compressions required for optimal generation of coronary perfusion pressure remains unknown although studies examining compression-to-ventilation ratios higher than 15:2 (C:V) in animals have reported higher C:V to be superior for return of spontaneous circulation and neurologic outcome. We examined human performance of two-rescuer CPR using various C:V. METHODS Thirty six EMT-Basic students in their final week of training performed two-rescuer CPR using C:V of 15:2, 30:2, 40:2, 50:2, and 60:2 on a recording resuscitation manikin. Compression and ventilation variables were recorded by computer while the number of pauses for ventilations and the hands-off time (time not spent performing chest compressions) were abstracted by hand. Data were analyzed by ANOVA and significant differences from the standard treatment of C:V = 15:2 were assessed by Tukey's HSD post hoc test. FINDINGS The number of compressions delivered per minute increased with increasing C:V while the hands-off time and pauses for ventilations decreased. All comparisons were significantly different from C:V = 15:2 (P < 0.001). The ventilation numbers decreased with increasing C:V although mean minute volume exceeded 1l for all C:V. INTERPRETATION A 15:2 compression-to-ventilation ratio when performed during two-rescuer CPR results in 26s of hands off time each minute while only delivering 60 compressions. Alternative C:V ratios of 30:2, 40:2, 50:2, and 60:2 all exceed the AHA recommended 80 compressions/min while still delivering a minute volume in excess of 1l.