Genetic predisposition and nongenetic risk factors of thiazolidinedione-related edema in patients with type 2 diabetes.

This study aimed to analyze the association of thiazolidinedione (TZD)-related edema with genetic and clinical variables and develop a simple points system to predict the risk of developing TZD-related edema. Fifty-eight (21.6%) of 268 individuals who received TZD for type 2 diabetes developed peripheral edema. Twenty-eight tag single nucleotide polymorphisms (SNPs) from candidate genes involved in sodium and water reabsorption were genotyped. Cox regression and logistic regression models were used to analyze the associations of different genotypes and weighted genotypic scores with TZD-related edema risk. Individuals with edema were older, predominantly female, and had greater weight gain. The AQP2 rs296766 T allele was associated with TZD-related edema [allelic P=0.0059; odds ratio (OR), 2.89; 95% confidence interval (CI), 1.61-5.17]. The SLC12A rs12904216 G allele had borderline significance (allelic P=0.049), which disappeared after correction for multiple testing. Patients with two SNP-based (AQP2 rs296766 and SLC12A1 rs12904216), who weighted genotypic scores within the top quartile, had a higher risk of developing TZD-related edema (OR, 16.45; 95% CI, 3.05-88.76). Combining the weighted genetic scores of two SNPs or all SNPs with age and sex information significantly improved the predictive power for TZD-related edema. We also developed a simple risk factor-based points system to predict an individual's risk of developing TZD-related edema. A clinically applicable prediction model including age, sex, and genetic information from AQP2 rs296766 and/or SLC12A rs12904216 SNPs can be developed to estimate the risk of TZD-related edema in type 2 diabetes patients.