OBJECTIVE--To see whether intravenous acetylcysteine would improve outcome in patients with fulminant hepatic failure after paracetamol overdose. DESIGN--A prospective randomised controlled study. SETTING--The Institute of Liver Studies, King's College Hospital, London. PATIENTS--50 consecutive patients (21 male) aged 16-60 with fulminant hepatic failure after paracetamol overdose who had not previously received acetylcysteine. INTERVENTIONS--Conventional intensive liver care plus either acetylcysteine (25 patients) in the same dose regimen as used early after a paracetamol overdose, except that the infusion was continued until recovery from encephalopathy or death, or an equivalent volume of 5% dextrose (25 patients). MAIN OUTCOME MEASURES--Survival; incidence of cerebral oedema, renal failure, and hypotension requiring inotropic support; liver function as assessed by prolongation of the prothrombin time; and degree of encephalopathy. RESULTS--The rate of survival was significantly higher in the acetylcysteine treated group than in the controls (48% (12/25 patients) v 20% (5/25); p = 0.037, 95% confidence interval for difference in proportions surviving 3% to 53%). Acetylcysteine treated patients had a lower incidence of cerebral oedema (40% (10/25) v 68% (17/25); p = 0.047, 95% confidence interval for difference in incidence 2% to 54%), and fewer developed hypotension requiring inotropic support (48% (12/25) v 80% (20/25); p = 0.018, 95% confidence interval 7% to 57%). Rates of deterioration and recovery of liver function, however, were similar in the two groups. No adverse reactions to acetylcysteine were seen. CONCLUSIONS--Acetylcysteine is safe and effective in fulminant hepatic failure after paracetamol overdose.