This single-center study examines the potential relationship between bystander knowledge of cardiopulmonary resuscitation (CPR) and in-hospital mortality rate among victims of out-of-hospital cardiac arrest (OHCA).246 patients were admitted to specialist units in our center with OHCA determined as being of cardiac etiology, between January 1990 and September 1993. We noted the presenting arrest rhythm, administration of bystander CPR, and the person giving CPR (family member, family doctor, off-duty doctor/nurse (D/N) and others (for whom CPR skill unknown). These factors were assessed in relation to in-hospital mortality.219 patients presented with ventricular fibrillation and 27 patients presented with electromechanical dissociation or asystole. 30 events were witnessed by ambulance crews, with 4 in-hospital deaths (13%). Of non crew-witnessed arrests, 108 patients received immediate CPR (55 deaths, 43%). 91 patients had delayed or no CPR (61 deaths, 70%). The in-hospital mortality difference is significant (χ2=5.27. 1DF, p<0.05). Mortality rates by resuscitators are shown below.ResuscitatorCasesDeathsMortalitySignificanceFamily doctor22836.4Off duty D/N25832Family member221568Other392461.5χ2=9.83, 3DF, p<0.05Administration of immediate CPR correlates with reduced in-hospital mortality offer OHCA. Patients attended by resuscitators with knowledge of CPR skills have significantly lower in-hospital mortality rates, even when the resuscitator is present by chance (off duty D/N) rather than attending because of prodromal symptoms (family doctor). Patients receiving CPR from family members had similar mortality to those receiving delayed or no CPR. Patients arresting in the presence of an ambulance crew have very low in-hospital mortality rates and were treated as a separate group.