OBJECTIVE: To assess the actual impact on coronary mortality of equipping ambulances with defibrillators. DESIGN: Retrospective analysis of routine medical and legal records of all those who had a cardiac arrest attributed to coronary heart disease occurring outside hospital in a defined population before and after the introduction of Heartstart. SETTING: City of Glasgow, North of the River Clyde, 1984 and 1990. PATIENTS: 296 and 267 men and women aged 25-64 inclusive in 1984 and 1990 respectively who had a cardiac arrest outside hospital which was attributed to coronary heart disease (International Classification of Diseases codes 410-414, ninth revision). RESULTS: The impact on coronary mortality in 1990 of equipping ambulances with defibrillators concurred with the earlier prediction of less than 1% of all coronary deaths. The circumstances of cardiac arrest were largely unchanged; most occurred outside hospital in the victim's home and the principal witnesses were members of the victim's family. A call for help before cardiac arrest was made in very few cases and cardiopulmonary resuscitation was attempted by laypersons in less than a third of the deaths they witnessed. There was a significant increase in the number of cardiopulmonary resuscitation attempts made by ambulance crews (16% v 32%, P < 0.01). Ambulance crews, however, still attended less than half of all cases (44% and 47%). CONCLUSION: The impact of equipping ambulances with defibrillators will remain small unless strategies are introduced that focus on improving the public's response to coronary emergencies by calling for help promptly and initiating cardiopulmonary resuscitation before the arrival of the emergency services.