Population Pharmacokinetic (Pop-PK) Analysis of Torsemide in Healthy Korean Males Considering CYP2C9 and OATP1B1 Genetic Polymorphisms

Abstract

Torsemide is a diuretic drug used for several cardiovascular and chronic diseases. With regard to the clinical application of torsemide, studies on individualized pharmacotherapy and modeling that take variability in pharmacokinetics (PKs) within a population into account have been rarely reported. Thus, the objective of this study was to perform population pharmacokinetic (Pop-PK) modeling and to identify effective covariates that could explain the inter-individual variability (IIV) of torsemide PK. Pop-PK modeling for torsemide was performed based on serum concentration data obtained from 112 healthy Korean males and analysis of various genetic and physicochemical parameters. Modeling was performed with nonlinear mixed-effects (NLME) using Phoenix NLME. The finally developed model was fully verified. The model was also reconfirmed using NONMEM software. As a basic model, the PKs of torsemide within the population were well described by a two-compartment model reflecting the lag-time on oral absorption. According to the genetic polymorphisms of OATP1B1 and CYP2C9, significant associations were found in the V/F, CL/F, and CL2/F of torsemide. These were reflected as effective covariates in the final Pop-PK model of torsemide, resulting in an approximately 5–10% improvement in the model parameter IIV values. Considering that torsemide is a substrate for CYP2C9 and OATP1B1, it was important to search for genetic polymorphisms in CYP2C9 and OATP1B1 as covariates to explain the PK diversity of torsemide between individuals. The differences in CL/F and CL2/F between the phenotypes of CYP2C9 were approximately 36.5–51%. The difference in V/F between the phenotypes of OATP1B1 was approximately 41–64.6%. These results suggested that the phenotypes of CYP2C9 and OATP1B1 produced significant differences in torsemide PKs. Considering that CYP2C9 and OATP1B1 phenotypes as covariates affected different PK parameters of torsemide, it could be inferred that torsemide’s cell membrane permeation process by OATP1B1 and the metabolic process by CYP2C9 could independently affect each other in vivo without interplay. There was no significant difference in the parameter estimates between modeling software (Phoenix NLME vs. NONMEM). In this study, the torsemide PK variability between individuals was largely explained. In the future, individualized effective drug therapy of torsemide taking individual patient’s genotypes into account might become possible.