In a region with a population of 250,000 people, all emergency calls for cardiac arrest were prospectively registered during a period of 6 years. Timing of events were carefully registered as were treatment and the participation of 3 ambulances equipped with defibrillators. When time until initial treatment of cardiac arrest was below 5 min, 12% could be resuscitated and discharged alive. This figure decreased to 2% in the period between 5 and 10 min and was zero to above 10 min. Similarly, a reasonable 12% of patients experiencing ventricular fibrillation at a public place could be resuscitated and discharged alive whereas only 5% of ventricular fibrillation occurring at the patients home could be successfully resuscitated. Asystolia was rarely treated successfully. Faster treatment improved results much and 63% of patients having ventricular fibrillation in the emergency room left hospital alive. Results of cardioversion in ambulances did not depend on time from initiation of cardiac arrest, but all patients receiving cardioversion later than 10 min died without regaining consciousness. The results were compared with other more effective programs. The study region apparently had much fewer cardiac arrest than a similar region in Seattle, U.S.A. In those cases where treatment could be initiated within 5 min, results were comparable.