Background The 2010 international guidelines for CPR recently recommended an increase in the minimum compression depth from 38 to 50 mm, although there are limited human data to support this. We sought to study patterns of CPR compression depth and their associations with patient outcomes in out-of-hospital cardiac arrest (OOHCA) cases treated by the 2005 guideline standards. Methods and Results We studied emergency medical services treated OOHCA patients from the Resuscitation Outcomes Consortium Epistry - Cardiac Arrest for whom electronic CPR compression depth data were available, from May 2006 to June 2009. We calculated anterior chest wall depression in millimeters and the period of active CPR (chest compression fraction) for each minute of CPR. We controlled for covariates including compression rate and calculated adjusted odds ratios for any return of spontaneous circulation (ROSC), 1-day survival, and hospital discharge. We included 1,029 adult patients from 7 U.S. and Canadian cities with these characteristics: mean age 68 years; male 62 %; bystander witnessed 40%; bystander CPR 37%; initial rhythms - VF/VT 24%, PEA 16%, asystole 48%, other non-shockable 12%; outcomes - ROSC 26%, 1-day survival 18%, discharge 5%. For all patients, median compression rate was 106 per minute, median compression fraction 0.65, and median compression depth 37.3 mm with 52.8% of cases having depth < 38 mm and 91.6% having depth < 50 mm. We found an inverse association between depth and compression rate (P<0.001). Adjusted odds ratios for all depth measures (mean values, categories, and in range) showed strong trends towards better outcomes with increased depth for all three survival measures. Conclusions We found suboptimal compression depth in half of patients by 2005 guideline standards and almost all by 2010 standards, as well as an inverse association between compression depth and rate. We found a strong association between survival outcomes and increased compression depth but no clear evidence to support or refute the 2010 recommendations of > 50 mm. While compression depth is an important component of CPR and should be measured routinely, the most effective depth is currently unknown.