Objective: Anticonvulsant drugs are known inducers of cytochrome P450 liver enzymes and it has been suggested that this induction increases susceptibility to paracetamol-induced hepatotoxicity. Methods: We measured the percentage urinary recovery of paracetamol and its metabolites after a dose of 20 mg kg−1, and the excretion of 6ß-hydroxycortisol as a ratio to urinary free cortisol(6ßOHF/F) in Chinese epileptic patients maintained on long term therapy with carbamazepine (n = 6) or phenytoin (n = 6). Results: Compared to the healthy controls (n = 20), patients on phenytoin had significantly lower recoveries of mercapturic acid, cysteine and sulphate metabolites, but a higher recovery of glucuronide metabolites of paracetamol. The recoveries of paracetamol metabolites in patients on carbamazepine were not different from controls. In contrast, the 6ßOHF/F was signi ficantly higher in patients on carbamazepine (3-fold) or phenytoin (2-fold) compared to controls. Healthy control Chinese subjects metabolised paracetamol in a similar way to that reported in Caucasians, indicating that the risk for hepatotoxicity would be the same. Our findings in a group of Chinese patients on phenytoin were also similar to those previously reported in Caucasians on this drug. The apparent differences in the pattern of isoenzyme induction between the groups on phenytoin and carbamazepine require verification in larger studies. The data do not suggest an increased risk of paracetamol-induced hepatotoxicity in Chinese patients on anticonvulsants.