The results of the first epidemiological, prospective, multicentric study on cardiac arrest in a geographical Italian region are reported. On 708 consecutive cardiac arrests, 438 underwent cardiopulmonary resuscitation (CPR). Of these, 344 were identified of cardiac aetiology. The underlying initial rhythm was: 166 asystole (48.3%), 104 ventricular fibrillation (30.2%), 74 pulseless electrical activity (21.5%). The best outcome occurred in patients whose cardiac arrest was witnessed by the EMS (49% return of spontaneous circulation (ROSC), 21% hospital discharge). When cardiac arrest was witnessed by lay people, 20.5% had ROSC and 4.4% were discharged alive from the hospital. When it was unwitnessed ROSC and hospital discharge were 8.6 and 1.7%, respectively. Ventricular fibrillation was highly predictive of outcome. Both ROSC and hospital discharge correlated inversely with the delay of the first defibrillation. Overall, the highest probability of survival was achieved when CPR interventions were started within the first minutes after collapse. Basic Life Support (BLS) manoeuvres began after 9 min of untreated cardiac arrest were still followed by a ROSC, but none of these patients survived. The incidence of prehospital cardiac arrest in our population was estimated to be in proportion of 0.95/1000 per year with a survival rate of 6.7%.