The pharmacokinetics of CS-023 (RO4908463, formerly R-115685), a novel parenteral carbapenem antibiotic, in humans was successfully predicted using the data collected from mice, rats, rabbits, and dogs; while inclusion of the monkey data led to a significant underestimation of the total plasma clearance (CL). Double logarithmic plots of CL and distribution volume at the steady-state (Vss) vs. body weight in four animal species were linear with high correlation coefficients; and the predicted CL and Vss values in humans agreed well with the observed values after administration of CS-023 by an intravenous drip infusion for 30 min. The plasma concentration-time profile in humans, which was predicted using a bi-exponential equation fitted to a complex Dedrick plot of the animal data, approximated the observed profile. An underestimation of CL caused by including the monkey data in a prediction is quite likely due to the net tubular reabsorption in monkeys, but not at least in rabbits, dogs, and humans.