Background: The effectiveness of epinephrine administration for cardiac arrests has been shown in animal models, but the clinical effect is still controversial. Methods and Results: A prospective, population-based, observational study in Osaka involved consecutive out-of-hospital cardiac arrest (OHCA) patients from January 2007 through December 2009. We evaluated the outcomes among adult non-traumatic bystander-witnessed OHCA patients for whom the local protocol directed the emergency medical service personnel to administer epinephrine. After stratifying by first documented cardiac rhythm, outcomes were compared among the following groups: non-administration, ≤10, 11-20 and ≥21min as the time from emergency call to epinephrine administration. A total of 3,161 patients were eligible for our analyses, among whom 1,013 (32.0%) actually received epinephrine. The epinephrine group had a significantly lower rate of neurologically intact 1-month survival than the non-epinephrine group (4.1% vs. 6.1%, P=0.028). In cases of ventricular fibrillation (VF) arrest, patients in the early epinephrine group who received epinephrine administration within 10min had a significantly higher rate of neurologically intact 1-month survival compared with the non-epinephrine group (66.7% vs. 24.9%), though other epinephrine groups did not. In cases of non-VF arrest, the rate of neurologically intact 1-month survival was low, irrespective of epinephrine administration. Conclusions: The effectiveness of epinephrine after OHCA depends on the time of administration. When epinephrine is administered in the early phase, there is an improvement in neurological outcome from OHCA with VF.  (Circ J 2012; 76: 1639–1645)