Abstract Objectives This study aimed to evaluate the effects of torsemide on warfarin therapy in humans and rats. Methods For the animal study, rats were orally dosed with warfarin (0.13 mg/kg, control group) or warfarin (0.13 mg/kg) with torsemide (2 mg/kg, low dose group and 10 mg/kg, high dose group). The pharmacodynamic response of warfarin was assessed by measuring the international normalized ratio (INR) for 5 consecutive days following drug administration. For the human study, 191 patients on warfarin with mechanical heart valves were followed up retrospectively. The stable dose was calculated as the mean dose in INR levels of 2–3 for 3 consecutive times. Key findings In the animal study, the INR, maximum plasma concentration (Cmax) and area under the plasma drug concentration–time curve (AUC0–∞) of (S)-warfarin in the high dose group were significantly higher than in other groups (P < 0.05). Compared with the control group, Cmax and AUC0–∞ of (R)-warfarin in the high and low dose groups were higher, whereas the volume of distribution/bioavailability and clearance/bioavailability were significantly lower (P < 0.05). In the univariate analysis of the clinical study, diuretics significantly lowered stable warfarin doses (P = 0.016) (5.07 ± 1.78 mg/day vs 5.77 ± 1.81 mg/day). After controlling confounding variables, the effects of diuretics were found to lower the warfarin dose by 0.464 mg. Conclusions It was concluded that warfarin dose needs to be lowered when it is used concomitantly with diuretics.