Aspects of the pharmacokinetics of warfarin that are clinically relevant are reviewed here. Since warfarin is normally completely absorbed, resistance to treatment due to impaired absorption is unusual, even in severe short bowel syndrome. Warfarin is highly albumin-bound; thus, hypoalbuminaemic states result in an increased free fraction of the drug and a decreased half-life but, as might be expected, there is no evidence of altered response at steady-state. Warfarin is completely metabolised by the liver to hydroxy-warfarins and warfarin alcohols, and although the latter have some biological activity they do not contribute significantly to the drug effect. No information is available concerning the metabolism of warfarin in chronic liver disease, but there is evidence of increased sensitivity due to impaired vitamin K-dependent clotting factor synthesis. Impaired renal function does not appear to alter the effect of warfarin. Lowered response to the drug may be secondary to poor compliance, kinetic resistance or pharmacodynamic resistance. These factors can be identified using algorithms based on population values for plasma warfarin concentrations and clearances at steady-state. The pharmacokinetics and pharmacodynamics of warfarin indicate that several days' overlap with heparin on initiation of warfarin, and gradual (rather than sudden) discontinuation of warfarin, might theoretically be necessary. However, those studies which have been performed have indicated that a long overlap and gradual discontinuation are not associated with greater safety or efficacy of the drug. Because of the long elimination half-life of warfarin and the short elimination half-life of vitamin K, many days' treatment with phytomenadione may be required after warfarin overdose. The elimination half-life and therefore the duration of therapy may be reduced by regular oral cholestyramine, although the means by which the latter enhances warfarin elimination is still unknown.