The chance of survival from ventricular fibrillation (VF) is up to ten times higher than those with other cardiac arrest rhythms. To calculate the effect of out-of-hospital resuscitation organisations on survival, it is necessary to know the percentage of cardiac arrest patients initially in VF and the relationship between delay time to defibrillation and survival. AIM: To study the incidence of VF at the time of cardiac arrest and on first ECG, the duration of VF and the relation between time to defibrillation and survival. METHOD: The Swedish Cardiac Arrest Registry has collected standardised reports on out-of-hospital cardiac arrests from ambulance organisations in Sweden, serving 60% of the Swedish population. RESULTS: In 14065 cases of out-of-hospital cardiac arrest collected between 1990 and 1995, resuscitation was attempted in 10966 cases. INCIDENCE: The first ECG showed VF in 43% of all patients. The incidence of VF at the time of cardiac arrest was estimated to be 60-70% in all patients and 80-85% in the cases with probable heart disease. DURATION: The estimated disappearance rate of VF was slow. Thirty minutes after collapse approximately 40% of the patients were in VF. SURVIVAL: Overall survival to 1 month was only 1.6% for patients with non-shockable rhythms and 9.5% for patients found in VF. With increasing time to defibrillation, the survival rate fell rapidly from approximately 50% with a minimal delay to 5% at 15 min. CONCLUSIONS: This study suggests a high initial incidence of VF among out-of-hospital cardiac arrest patients and a slow rate of transformation into a non-shockable rhythm. The survival rate with very short delay times to defibrillation was approximately 50%, but decreased rapidly as the delay increased.