AIMS: To describe changes in different factors at resuscitation and survival in a 17-year survey of patients suffering from out-of-hospital cardiac arrest. METHOD: The investigation was carried out in the community of Göteborg with 450 000 inhabitants during 1981-1997 on all patients suffering out-of-hospital cardiac arrest in whom resuscitation was attempted. RESULTS: The number of cases per year, the proportion of witnessed arrests and the proportion of arrests of cardiac aetiology remained similar over time. There was an increase in median age from 68 to 73 years (P<0.0001), in the proportion of females from 27% to 33% (P=0.035) and in the proportion of patients receiving bystander cardiopulmonary resuscitation from 14% to 28% (P<0.0001) with time. There was a shortening of the median interval from collapse until defibrillation from 9 min to 6 min (P<0.0001) over time but a decrease in the occurrence of ventricular fibrillation as the initially recorded arrhythmia from 39% to 32% (P=0.022). There was an increase in the proportion of patients having a bystander witnessed cardiac arrest of cardiac aetiology being hospitalized alive from 32% to 45% (P<0. 0001 for change over time). The proportion of patients discharged alive from hospital increased from 16% to 29% until 1993, but thereafter decreased to 13% in 1997 (P=0.002 for change over time). CONCLUSION: In a survey covering 17 years of resuscitation of out-of-hospital cardiac arrest patients we found that the occurrence of ventricular fibrillation as the initially recorded arrhythmia decreased. There was an increase in age, in the proportion of females and in the use of bystander cardiopulmonary resuscitation. The interval between collapse and defibrillation was shortened. Survival changed over time with an increase until 1993 but with a decrease thereafter.