Combined Influence of Genetic Variants and Gene-gene Interaction on Sulfonylurea Efficacy in Type 2 Diabetic Patients

Several genes have been shown to influence the response to sulfonylureas in previous studies. However, their interactions and combined genetic influence on sulfonylurea efficacy remain mostly unexplored. So the aim of this study was to examine the interactions of these published susceptibility alleles and access the potential utility of combining information for drug response prediction.

We genotyped 5 variants of these genes in a total of 747 diabetic patients enrolled in a trial of glibenclamide (CYP2C9, KCNJ11, TCF7L2, KCNQ1 and CDKN2A/2B gene). All the patients were followed for 48 weeks. Treatment failure was fasting blood glucose level≥7.0 mmol/L. A Cox regression model was used to evaluate the relationship between genetic variants and treatment failure over a period of 48 weeks. The receiver operating characteristic (ROC) curve, the area under the curve (AUC) and the multifactor dimensionality reduction method (MDR) analyses were used to assess the gene-gene interaction and also the predictive power of the combined variants.

The relationship beween risk alleles and FBG reduction at the end of the first month was not significant (Low Risk Group 10.8% (1.2-20.7%), Middle Risk Group 10.7% (1.3-21.1%), High Risk Group 8.51% (- 0.1-18.4%), P=0.212). After adjusting for sex, age, BMI, total dose of glibenclamide and baseline HbA1c, Cox regression showed a borderline significant association between the number of risk alleles and glibenclamide treatment failure (HR=1.125, 95%CI 1.016-1.246, P=0.023). There was potential gene-gene interaction between KCNJ11 and CDKN2A/2B gene using MDR method. The area under the ROC curve was 0.547, and the testing balanced accuracy of the best genetic model in MDR analysis was 0.5643.

Our study demonstrated that the combined genetic variants were borderline significantly associated with the efficacy of glibenclamide, and there are gene-gene interaction between KCNJ11 and CDKN2A/2B. More evidence was needed for increasing the predictive value of genetic variants on glibenclamide efficacy.