The mechanisms of the reported dose-dependent warfarin pharmacokinetics were investigated using [14C]warfarin. When administered in microdoses (9 micrograms i.v.) to rats (male Wistars, 270-300 g), a steep distribution phase (T1/2 = 0.25 hr) was followed by a relatively slow beta-phase (T1/2 = 40 hr). The observed volume of distribution was 390 ml. This pharmacokinetic behavior contrasted highly with the one seen for higher (greater than 0.2 mg/kg) doses (unlabeled) warfarin; volume of distribution = 45 ml, T1/2 = 12.5 hr. If a "macrodose" (0.2 mg/kg) preceded (16 hr) the "microdose," "normal" pharmacokinetics were observed for the latter, suggesting a saturable "deep compartment." The administration of 4-hydroxycoumarins (i.e., acenocoumarol, phenprocoumon and warfarin) after the microdose of [14C]warfarin was in its beta-phase caused a rapid rise of plasma [14C]warfarin indicating [14C]warfarin to be displaced from the "deep compartment." The rate of appearance of [14C]warfarin was 0.3 hr-1 irrespective the 4-hydroxycoumarin used. The hepatic distribution of [14C]warfarin was investigated and the effect of a displacer thereupon. Fifty-three hours after the [14C]warfarin administration, the liver contained about 40% of the dose; 45% of it was bound to microsomes. The administration of acenocoumarol (0.2 mg/kg) at 48 hr, halved the liver content. [14C]warfarin was redistributed from microsomes (-65%) and from the 10,000 X g pellet (-65%) into the cytosol (+260%) and the plasma (+320%). Microsomal bound [14C]warfarin in vitro could not be washed out or be displaced unless dithiothreitol (50 mM) was included in the washing buffers.(ABSTRACT TRUNCATED AT 250 WORDS)