Current resuscitation guidelines recommend that defibrillation be undertaken as soon as possible in patients suffering a cardiac arrest where the cardiac rhythm is either ventricular fibrillation (VF) or ventricular tachycardia (VT). Evidence from animal and clinical studies suggests that outcomes may be improved if a period of cardiopulmonary resuscitation (CPR) is given prior to defibrillation. The objective of this study was to determine if 90 seconds of CPR before defibrillation improved survival.Patients suffering non-paramedic witnessed VF/VT cardiac arrest were randomized to receive either 90 seconds of CPR before defibrillation (treatment) or immediate defibrillation (control). The study was carried out in Perth, Western Australia between June 2000 and June 2002. The primary endpoint was survival to hospital discharge with secondary endpoints of return of spontaneous circulation (ROSC) and survival at 1 year.A total of 256 patients underwent randomization. Baseline characteristics including response intervals were similar in both groups. Survival to hospital discharge in the CPR first group was 4.2% (5/119) compared with 5.1% (7/137) for the immediate defibrillation group (OR 0.81; 95%CI. 0.25-2.64). No difference in those achieving ROSC was observed between the groups (OR 1.16; 95% CI 0.49-2.80).Ninety seconds of CPR before defibrillation does not improve overall survival in patients suffering VF/VT cardiac arrests. Further studies to evaluate various aspects of this treatment strategy are required as published outcomes to date are inconclusive.