This investigation was designed to determine if the reported stereoselectivity of the pharmacokinetic interaction between warfarin and metronidazole in humans occurs also in rats and if the potentiation of the anticoagulant effect of warfarin can be ascribed solely to inhibition of warfarin metabolism by metronidazole. Metronidazole had no effect on the serum protein binding of racemic warfarin in vitro over a wide concentration range but decreased the protein binding of R-(+)-warfarin and S-(-)-warfarin in vivo, perhaps indirectly through metabolite(s). Treatment with i.p. metronidazole, 100 mg/kg every 6 hr, decreased the plasma clearance of free warfarin ("intrinsic clearance"). This inhibitory effect was more pronounced with S-(-)-warfarin than with R-(+)-warfarin (congruent to 60 and congruent to 30%, respectively). Metronidazole did not affect plasma prothrombin complex activity in vitro but reduced it in vivo. Metronidazole treatment increased the elimination rate constant for endogenous prothrombin complex activity and decreased the plasma concentration of free R-(+)-warfarin required to decrease prothrombin complex activity synthesis rate to one-half of normal (there were insufficient data to determine the effect of metronidazole on the activity of the other enantiomer). It is concluded that metronidazole preferentially inhibits the metabolism of S-(-)-warfarin in rats as in humans but that (at least in rats) the stereoselectivity is not absolute. Moreover, metronidazole has an indirect inhibitory effect on warfarin protein binding in vivo and also affects the pharmacodynamics of warfarin as well as the blood clotting process per se.