Aim: Cardiac arrest with ventricular fibrillation (VF) has been divided into three phases in which phase-specific therapy may improve outcome. The aim of this study was to assess the relationship between call-to-shock time, bystander CPR (BCPR), and cardiac arrest outcomes. Methods: In a retrospective analysis of prospectively-acquired data from witnessed VF out-of-hospital cardiac arrests (OHCA), patients were classified as phases 1, 2, or 3 based on call-to-shock time (<5, 5-8, and >8 min) and further stratified based on performance of BCPR. Groups were compared with regard to survival, neurological outcome, and restoration of spontaneous circulation (ROSC) with defibrillation only (no ALS interventions to achieve sustained ROSC). Results: Survival, neurologically intact survival, and ROSC with defibrillation were different between phases 1 and 2 (p = 0.014 and p = 0.005, p < 0.01) but not between phases 2 and 3. Patients were further classified as having received BCPR (N = 111) or no BCPR (N = 107). Neurologically intact survival with and without BCPR, respectively, was 61% versus 72% (phase 1), 44% versus 41% (phase 2), and 42% versus 29% (phase 3). ROSC with defibrillation only with and without BCPR, respectively, was 64% versus 56% (phase 1), 37.0% versus 29% (phase 2), and 33% versus 8% (phase 3). ROSC with defibrillation alone was statistically higher in univariate analysis in phase 3 with BCPR (p = 0.033) but not in multivariate analysis (p = 0.068). Conclusions: BCPR did not significantly improve survival in any phase of OHCA, though there was a trend toward increased neurologically intact survival and increased ROSC with defibrillation alone in phase 3. © 2006 Elsevier Ireland Ltd. All rights reserved.