OBJECTIVE: To determine the epidemiology of out-of-hospital cardiac arrests and survival after resuscitation and to apply the Utstein style of reporting to data collection. DESIGN: Prospective cohort study. SETTING: A middle-sized urban city (population 516,000) served by a single emergency medical services system. PATIENTS: Consecutive prehospital cardiac arrests occurring between 1 January and 31 December 1994. INTERVENTION: Advanced cardiac life support according to the recommendations of American Heart Association. MAIN OUTCOME MEASURES: Survival from cardiac arrest to hospital discharge, and factors associated with survival. RESULTS: Four hundred and twelve patients were considered for resuscitation. The overall incidence of out-of-hospital cardiac arrest was 79.8/100,000 inhabitants/year. Fifty seven patients (16.6%) survived to discharge when resuscitation was attempted. 32.5% survived when cardiac arrest was bystander witnessed and was of cardiac origin with ventricular fibrillation as the initial rhythm. When asystole or pulseless electrical activity was the first rhythm recorded, discharge rates were 6.2 and 2.7% respectively. The cause of cardiac arrest was cardiac in 66.5%, and ventricular fibrillation was the initial rhythm in 65.0% of bystander witnessed cardiac arrests of cardiac origin. 22.1% of patients received bystander initiated cardiopulmonary resuscitation. The mean time intervals from the receipt of the call to the arrival of a first response advanced life support unit and mobile intensive care unit at the patient's side and to the return of spontaneous circulation were 7.0 and 10.3 and 12.6 and 16.7 min respectively. In the logistic regression model bystander witnessed arrest, age, ventricular fibrillation as initial rhythm, and the call-to-arrival interval of the first response unit were independent factors relating to survival. Utstein style reporting with modification of time zero was found to be an appropriate form of data collection in this emergency medical services system. CONCLUSIONS: After implementation of major changes in the emergency medical services system during the 1980s survival from out-of-hospital cardiac arrest markedly increased. However, early access, which has turned out to be the weakest link in the chain of survival, should receive major attention in the near future. Utstein style reporting with a modified time zero was found to be appropriate, although laborious, protocol for data collection.