The objective of our study was to characterise 52 hospital survivors of pre-hospital ventricular fibrillation and record their initial management in hospital. A retrospective review was undertaken of ambulance report forms, hospital notes, and electrocardiograms, in one teaching hospital and three district general hospitals in South Wales, of 53 patients discharged from hospital between February 1987 and April 1992 after resuscitation from pre-hospital ventricular fibrillation by ambulance personnel. Twenty patients showed evidence of acute myocardial infarction (group 1), eight patients had a diagnosis of 'possible acute myocardial infarction' (group 2), and 25 patients had no evidence of acute myocardial infarction (group 3). Nineteen patients in group 1 experienced chest pain before collapse compared with only six patients in group 3 (p < 0.001). Five patients in group 1 had a previous history of ischaemic heart disease compared with 17 patients in group 3 (p < 0.01). A greater proportion of patients in group 3 were taking diuretic medication (15 of 25 vs 4 of 20: p < 0.01) but there was no difference in potassium levels on admission to hospital. Cardiologists were involved in the management of a minority of patients (21 of 53); only eight patients underwent cardiac catheterisation; and only three were referred for electrophysiological studies. Patients in group 3 were more likely to be discharged taking empiric antiarrhythmic drugs (13 of 25) than patients in group 1 (2 of 20) (p < 0.01). Not enough use is made of noninvasive and invasive investigations in the management of survivors of pre-hospital ventricular fibrillation.(ABSTRACT TRUNCATED AT 250 WORDS)