OBJECTIVE: To identify the impact of advanced life support skills on outcome for prehospital cardiac arrest in a defined population and to assess the value of certain physiological variables in predicting the outcome in those successfully resuscitated in the accident and emergency (A&E) department; to identify areas for improvement in the outcome of such patients. DESIGN: Prospective 12 month study. SETTING: Leicestershire, United Kingdom. MAIN OUTCOME MEASURE: Survival to hospital discharge and status at 6 months. RESULTS: 266 patients were identified as having suffered a prehospital cardiac arrest; of these, 86 had their resuscitation attempt terminated in the community by a general practitioner and 180 were transferred to the A&E department of the Leicester Royal Infirmary. Of the latter, 159 were felt to be of cardiac aetiology, and 19 were eventually discharged from hospital. All survivors had experienced a witnessed cardiac arrest, ventricular fibrillation (VF) being identified as the initial rhythm. After adjusting for age and sex using logistic regression, the Glasgow coma score (GCS) was found to be associated with subsequent mortality (chi 2 = 18.22 on 2 df, P < 0.0001). Compared to a baseline GCS of 9-15, the relative odds of death for a GCS of 3 were 25.3 (95% confidence interval 4.3 to 149-9), while a GCS of 4-8 gave a relative odds of death of 12-18 (95% CI 1.8 to 80.2). No significant association was found between postarrest arterial pH and mortality. CONCLUSIONS: The immediate GCS on admission is a predictor of outcome and it is important to monitor its trend in the first 24 h. Multidisciplinary audit and joint guidelines with other specialties are important in optimising the care of these patients.