The basis for susceptibility to halothane-induced liver necrosis in guinea-pigs was examined. In hepatic microsomes, the following were similar in susceptible and resistant animals: total cytochrome (CYP) P450 (P450), phenobarbital-inducible pathways of mixed function oxidation (androstenedione 6 beta- and 16 beta-hydroxylase activities) and the CyP2E1-catalysed pathway of N-nitrosodimethylamine N-demethylase activity. Similarly, immunohistochemical staining of CYP2E1 protein was equivalent in livers from susceptible and resistant guinea-pigs. Prior treatment with the P450-inhibitors, metyrapone and SKF-525A ameliorated halothane-induced liver damage in susceptible animals. Conversely, in resistant guinea-pigs, stimulation of hepatic CYP2E1 activity by treatment with 4-methylpyrazole produced severe hepatotoxicity after re-exposure to halothane. These results confirm the conclusions of others, that P450-mediated metabolism produces halothane-induced liver necrosis in the guinea-pig model but, as in other work, the data fail to explain why no difference in activity of these enzymes could be found between susceptible and resistant guinea-pigs. To establish whether a differential effect on hepatic blood flow between susceptible and resistant guinea-pigs could explain this paradox, studies were performed using a radiolabelled microsphere technique. The effect of halothane on lowering cardiac output was identical in both groups of animals and halothane significantly reduced hepatic arterial but not portal blood flow. The effect on arterial blood flow was more profound in susceptible guinea-pigs (0.67 +/- 0.17% of injected microspheres) than in resistant animals (0.99 +/- 0.13%; P < 0.005). It is concluded that P450-catalysed metabolism and reduced hepatic blood flow are both necessary to produce halothane-induced liver injury in susceptible guinea-pigs, but it is the effect of halothane on hepatic arterial blood flow that differs between susceptible and resistant animals.