INTRODUCTION The 2010 AHA / ERC Guidelines recommend compression depths of 5-6 cm. However, the neutral sternal position is assumed to remain unchanged. Changes in chest wall or lungs might cause incomplete chest recoil, altering the effect of 5-6 cm as a percentage of chest diameter. This could induce injury and affect the optimal compression depth. We investigated the hypothesis that incomplete chest recoil occurs during CPR in a pig model and that the speed at which this occurs changes with prolonged CPR. The incomplete chest recoil, or chest molding is a process of change, not a limitation to re-expansion caused by “leaning” or partial compression relaxation. MATERIAL AND METHODS Fourteen Yorkshire pigs with an average AP diameter of 20 cm were anesthetized, placed in dorsal recumbency, and ventricular fibrillation (VF) induced. After 90 s of VF, a specially made, proprietary servo controlled automated CPR device was used to deliver compressions of 5 cm depth at various frequencies (80-120 cpm) and duty cycles (40-60%). The plunger position (and unstressed sternal position) was determined at the start of each compression by the servo system. The position of the compression plunger (and thorax) was measured at a sample frequency of 250 Hz. An altered position at the start of a new compression is called the molded position. RESULTS In the 14 animals, chest molding occurred with an average initial rate of 0.34 ± 0.12 mm per compression. The average chest molding after 100 compressions was 17.4 ± 3.6 mm. After 21 minutes the average molding was 29.2 ± 10.3 mm with an average molding rate of -0.02 ± 0.04 mm/compression. The moulding curve shape was logarithmic. Initial compression depth was approximately 25% of AP diameter, increasing to 30% after 21 minutes CPR due to chest molding. CONCLUSION In conclusion, the molding magnitude was substantial. After initiation of CPR, chest molding occurs quickly, the rate slowing but never stopping, until compressions are stopped. The results could be extrapolated to humans but may be less pronounced due to a more vertical rib construction. Implications on CPR quality and injury need further investigation. Nevertheless, chest molding should be considered in the instructions for compression depth.