A recent experience of the senior author when addressing a meeting of anaesthetists prompted this review, since it was apparent that there was still considerable disagreement between hepatologists and anaesthetists over the association between liver cell damage and halothane anaesthesia and the measures by which the risk could be minimised. Although during the 1960s and 1970s the evidence was hotly debated, there has since been increasing acceptance, both clinical and experimental, of a strong prima facie case for an association.'" Two, probably distinct, forms of liver damage have been defined.'2 Serum aminotransferase activities are raised in up to a fifth of patients anaesthetised with halothane during the first and second postoperative weeks (type I). Such minor forms of liver injury are to be distinguished from the rare occurrence of massive liver cell necrosis (type II).Three well controlled studies of minor reactions have been reported from Britain.s 6 8 Significant rises in serum aspartate aminotransferase (AST) activity during the postoperative period were found only in those patients receiving halothane,5 and these activities were significantly higher than in those receiving trichloroethylene.6 Abnormalities in liver function were not always apparent until the second postoperative week. Rises of serum amunotransterase activity were also greater in those receiving halothane than with enflurane,8 and obese women were more likely to develop abnormalities.Most patients who have developed massive necrosis have had a previous and milder reaction to halothane. Nevertheless, the frequency of minor abnormalities (up to 20%) and the very low incidence of massive necrosis make it clear that minor reactions are not necessarily followed by more severe effects. There is no way of predicting which patients will follow this course.We have seen now 48 patients with otherwise unexplained massive liver cell necrosis after halothane anaesthesia referred to the liver unit over the period January 1965 to December 1983. In each of the 48 patients other possible causes of liver damage were excluded, including exposure to hepatotoxic agents, sepsis, hypotension during surgery, and infection with hepatitis A and B, cytomegalovirus, and Epstein-Barr virus. In no case was there evidence of pre-existing liver disease.Thirty one of the 48 patients were women, giving a female to male ratio of 1 8:1. Ages ranged from 21 to 76 years (median 57 years), contrasting appreciably with the much younger age distribution of patients with fulminant viral hepatitis (fig 1). In comparison with the age distribution of patients undergoing anaesthesia in England and Wales, as given in the Hospital Inpatient Enquiry,'3 the patients with halothane hepatotoxicity are in a slightly older age group. Sixty eight per cent of the patients were obese and there was a history of allergy to other drugs in one third of them.The interval between the last exposure to halothane and the onset of jaundice ranged from two to 26 days, with a median of five days, and i...