Genetics of Metformin Response in the Kaiser Permanente Northern California (KPNC) Diabetes Registry

Trial description: Metformin is a commonly used first-line treatment for controlling type 2 diabetes mellitus (T2DM). Metformin’s effect on long term glucose levels is generally measured as a reduction in hemoglobin A1c (HbA1c) with reductions ranging from 1 to 2%. Despite Metformin’s frequent use the glycemic response varies with as many as 35% of subjects failing to achieve acceptable glycemic control. Extrinsic factors such as dosing influence response, however intrinsic factors including subject genetic profile may also play a role. The objective of this study is to conduct a genome wide association (GWA) study of Metformin response in good and poor responders to Metformin to identify genetic variants that influence Metformin response. To that end, we identified and genotyped subjects with T2DM from Kaiser Permanente Northern California Diabetes Registry, who were taking Metformin and evaluated their response to Metformin as on treatment HbA1c levels adjusted for baseline HbA1c, age, gender, average metformin dose and serum creatinine (further detailed in Selby JV, et al. Diabetes Care 1997;20:1396–402). We then conducted a GWA study on 524 good Metformin responders and 478 poor Metformin responders. Specifically, subjects were genotyped on the Illumina OmniExpressExome genome-wide array and imputation was done using the phase 3 1000 Genomes reference panel. Logistic regression analyses were conducted to test the effect of genotype on Metformin response while controlling for population stratification using principle components for ancestry and key covariates including age at diagnosis, body mass index (BMI), baseline HbA1c level and dosage. A genome wide level of significance was used (p<5x10^-8) and an a priori selected candidate variants were considered significant at a p<0.00026.