To quantify the effects of the liver on the dose dependence of plasma warfarin clearance, an equal number of normal and functionally hepatectomized rats received an intravenous bolus of either 0.01, 0.1, or 1.0 mg/kg body weight of radiolabelled sodium warfarin. Serial samples of plasma and bile collected from each rat during the 1 hr experiment and of hepatic tissue collected at the end of the experiment were analyzed for radioactivity. The disappearance of warfarin from the plasma of hepatectomized rats was not dose dependent and suggested that the apparent dose dependency of plasma warfarin clearance is primarily the result of warfarin's interaction with hepatic tissue. The disappearance of warfarin from the plasma of normal rats was dose dependent with higher doses being cleared less rapidly. This dose dependence, however, was not reflected in the rate of biliary excretion of warfarin's metabolites, which did not show saturation over this dosage range. These results were used to develop a multicompartmental model of warfarin's pharmacokinetics. Plasma warfarin data collected from the hepatectomized rats were used to develop the extrahepatic components of the model, which was then expanded to include hepatic tissues based on data collected from normal rats. To simultaneously fit the plasma, biliary, and hepatic data required that at least two classes of hepatic tissue exchange warfarin with plasma. One tissue exhibited Michaelis-Menten saturation kinetics with Kd and maximum capacity estimated at 1.49E - 3 micrograms/ml and 2.72 micrograms/ml, respectively. The second class exhibited linear exchange kinetics with free plasma warfarin. Warfarin's association with the second class of hepatic tissue leads to its metabolic elimination. Consistent with our experimental findings, the rate of warfarin elimination from the plasma into the bile was linearly related to plasma warfarin concentration. Thus the single hepatic exchange nonlinearly was necessary and sufficient to account for the apparent dose dependency in plasma warfarin's pharmacokinetics. These results suggest that over the range of doses studied, the apparent dose dependent differences in the plasma warfarin concentration profile can be accounted for by saturable hepatic uptake. This mechanism, however, is not related to warfarin's metabolic enzymes, which do not show saturation in the dosage range studied.