OBJECTIVE To elucidate the pharmacokinetics of amphotericin B in rats, mice and humans, and to perform interspecies scaling to humans using allometry. METHODS Plasma concentrations following intravenous bolus administration in rats, and mice were determined by HPLC. Human pharmacokinetic parameters elucidated from literature data were validated in a preliminary study involving a patient receiving daily infusion dose for 27 days. A critical literature review was conducted to identify appropriate pharmacokinetic parameter values in other species for interspecies scale-up. Interspecies allometric scale-up was performed across mice, rats, rabbits and dogs and the resulting predictions in humans were compared to observed values. RESULTS A triexponential decline in rat, mouse and human plasma concentrations were observed. No gender differences in rat pharmacokinetics were observed. In contrast to allometry, mouse CL was smaller (82 vs 116 ml/h/kg) and T0.5 (33 vs 20 h) was longer compared to rat. In the preliminary human study, Cpeak and Cmin values remained relatively constant over the duration of therapy, and a CL, MRT, T0.5, Vss and Vdarea of 26 ml/h/kg, 10 and 23 days, 6.2 and 20 L/kg, respectively, were estimated. The relative contributions of the terminal phase area in rat, mouse and human were 75%, 92% and 31%, respectively. Interspecies allometric scale-up predictions of human CL (41 ml/h/kg), CLu (467 ml/h/kg) and Vss (3.3 L/kg) were similar to reported values, whereas poor predictions of human Vuss (33 L/kg), Vdarea (4.1 L/kg) and T0.5 (3 days) were obtained. CONCLUSIONS Insignificant accumulation in humans inspite of the long terminal T0.5 was rationalized to be due to the small terminal-phase area contribution. While human CL and Vss were successfully predicted in the interspecies scaling, poor predictions of human Vdarea and T0.5 were obtained, which was attributed to disposition pattern differences between humans and other species, a potential new critical factor affecting interspecies scale-up.