To describe the epidemiology and survival from out-of-hospital cardiac arrest.Longitudinal follow-up study from the time of paramedic attendance to 12 months later.Perth, Western Australia (WA), a metropolitan capital city with an adult population of approximately one million people.The St John Ambulance Australia (WA Ambulance Service Incorporated) cardiac arrest database was linked to the WA hospital morbidity and mortality data using probabilistic matching.Of 3730 cardiorespiratory arrests in 1996-1999, the age standardised rate of arrests of presumed cardiac origin, where resuscitation was attempted (n=1293) was 32.9 per 100000 person-years and 7.1 per 100000 person-years for bystander-witnessed VF/VT arrests.Survival to 28 days was 6.8% following all bystander-witnessed cardiac arrests; 10.6% following bystander-witnessed VF/VT arrests and 33% for paramedic-witnessed cardiac arrests. Logistic regression analysis showed an inverse association between ambulance response time interval and survival following all bystander-witnessed cardiac arrests (and VF/VT arrests). ONE YEAR SURVIVAL: 89% of bystander-witnessed cardiac arrest survivors and 92% of paramedic-witnessed cardiac arrests were still alive at 1 year post-arrest.The trends in occurrence and survival following out-of-hospital cardiac arrest in Perth, WA, are similar to those found elsewhere. There is an opportunity to strengthen the chain of survival by reducing the response time interval and increasing the use of bystander cardiopulmonary resuscitation (CPR). First-responder programs and public access defibrillation will need to be considered in the light of local demographics, location and the epidemiologic features of out-of-hospital cardiac arrest.