This trial compared blinded 10 mg aliquots of adrenaline with placebo in 194 cardiac arrest patients treated in hospital using American Heart Association guidelines. In-hospital and out-of-hospital arrests were included. Of the 339 eligible patients a large proportion (145 (45%)) were not randomised and received open 1 mg aliquots of adrenaline. This group is also analysed. Supervising physicians gave significant preference for males, patients with no previous cardiac history and without multiple organ disease to be given open 1 mg adrenaline. Patients in asystole at the time of consideration for entry were preferentially placed in the trial group (114 (69%) vs. 170 (88%)) and patients in ventricular fibrillation were preferentially given open 1 mg adrenaline (31 (21%) vs. 24 (12%) P < 0.03). The most beneficial rhythm changes which led to survival were sinus rhythm and ventricular tachycardia. Analysis of rhythm changes resulting from the dosing showed a significant (P = 0.01) change to a beneficial rhythm with 10 mg adrenaline but not for 1 mg adrenaline or placebo. This was not reflected by an improvement in immediate survival. No significant differences in immediate survival (IS) or hospital discharge (HD) exists between open 1 mg adrenaline (IS 14 (9.7%), HD 3 (2%)) or the 10 mg adrenaline (IS 9 (9.6%), HD 0) vs. placebo (IS 7 (7%), HD 0) trial arms. Patients reaching the point of use of adrenaline have a uniformly poor immediate survival (8.8%) and hospital discharge rate (0.9%). Dosing with 10 mg or 1 mg adrenaline does not influence outcome compared with placebo.