The purpose of this analysis was to develop a population pharmacokinetic model for CS-917, an oral hypoglycemic prodrug, and its 3 metabolites. The population pharmacokinetic model was used to predict exposure of the active moiety R-125338 and thus to identify potential CS-917 dosage reduction criteria. The dataset included 6 phase I and IIa studies in patients with type 2 diabetes mellitus. The pharmacokinetic profile of CS-917 and its metabolites was described by a series of linked 1- and 2-compartmental models. Simulations showed that moderate renal impairment has a clinically significant impact on exposure to R-125338. A separate population pharmacokinetic analysis of R-125338 alone revealed similar results. In conclusion, a population pharmacokinetic model fit to the active moiety alone yielded similar predictions and substantially reduced the analysis time compared to the more complex model developed for CS-917 and its metabolites. Increased exposure to R-125338 in the presence of moderate renal impairment may be an important consideration for dose selection.