Population Pharmacokinetics of Metformin in Healthy Subjects and Patients with Type 2 Diabetes Mellitus: Simulation of Doses According to Renal Function Janna K. DuongShaun S. KumarCarl M. KirkpatrickLouise C. Greenup • Manit AroraToong C. LeePeter TimminsGarry G. GrahamTimothy J. Furlong • Jerry R. GreenfieldKenneth M. WilliamsRichard O. Day

Background and Objective Metformin is contraindicated in patients with renal impairment; however, there is poor adherence to current dosing guidelines. In addition, the pharmacokinetics of metformin in patients with significant renal impairment are not well described. The aims of this study were to investigate factors influencing the pharmacokinetic variability, including variant transporters, between healthy subjects and patients with type 2 diabetes mellitus (T2DM) and to simulate doses of metformin at varying stages of renal function. Methods Plasma concentrations of metformin were pooled from three studies: patients with T2DM (study A; n = 120), healthy Caucasian subjects (study B; n = 16) and healthy Malaysian subjects (study C; n = 169). A population pharmacokinetic model of metformin was developed using NONMEM version VI for both the immediate-release (IR) formulation and the extendedrelease (XR) formulation of metformin. Total body weight (TBW), lean body weight (LBW), creatinine clearance (CLCR; estimated using TBW and LBW) and 57 singlenucleotide polymorphisms (SNPs) of metformin transporters (OCT1, OCT2, OCT3, MATE1 and PMAT) were investigated as potential covariates. A nonparametric bootstrap (n = 1,000) was used to evaluate the final model. This model was used to simulate 1,000 concentration‐time profiles for doses of metformin at each stage of renal impairment to ensure metformin concentrations do not exceed 5 mg/l, the proposed upper limit. Results Creatinine clearance and TBW were clinically and statistically significant covariates with the apparent clearance and volume of distribution of metformin, respectively. None of the 57 SNPs in transporters of metformin were significant covariates. In contrast to previous