Despite early controversy, it is now recognised that halothane anaesthesia may be followed by abnormalities of liver function. The resulting hepatitis may take 1 of 2 forms: in type I, there is a minor degree of disturbance of liver function shown by increased serum transaminases or glutathione-S-transferase in up to 25 to 30% of patients; subsequent re-exposure to halothane is not necessarily associated with evidence of liver damage. In contrast, type II hepatitis is often associated with massive liver cell necrosis, frequently leading to fulminant hepatic failure. This type of liver damage has clinical, serological and immunological features compatible with an immune-mediated idiosyncratic reaction. The incidence is low (between 1 in 3500 and 1 in 35,000 anaesthetic procedures), but the mechanism of halothane hepatitis remains uncertain: there have been extensive animal models showing that halothane has a direct hepatotoxic potential, although the relevance of this to the human patient is not yet clear. Prevention of halothane hepatitis may be difficult, and the only clear way of reducing the incidence is to avoid re-exposure to halothane in those patients who have had a previous adverse reaction to the drug, demonstrated either by unexplained pyrexia or by jaundice. Halothane should also be avoided in those patients where there is a family history of sensitisation to the drug. In such cases, halothane-free equipment should be used, and exposure to other volatile non-halogenated anaesthetics should be avoided.