Genetic Associations with Type 2 Diabetes Differ According to Metabolic Contexts Defined by Insulin Resistance and Visceral Adiposity

Stratifying populations based on indicators of insulin resistance and visceral adiposity may help identify genetic associations with type 2 diabetes mellitus (T2DM) that vary across metabolic environments. This study aimed to investigate metabolic context-dependent genetic associations with T2DM using the metabolic score for insulin resistance (METS-IR) and visceral fat (METS-VF). A total of 70,583 and 70,513 participants were eligible for METS-IR and METS-VF stratification, respectively. After excluding participants with missing diabetes status, the final case–control samples comprised 40,970 and 40,944 participants, respectively. Genome-wide association studies of prevalent T2DM were conducted within metabolic tertiles, with additional evaluation of lead variants in the intermediate tertile (T2). Logistic regression under an additive genetic model was performed, followed by genotype × metabolic stratum interaction analyses and Cochran’s Q heterogeneity tests. Twelve notable loci were identified, showing both shared and varying association patterns across metabolic strata. CDKAL1 showed robust associations across metabolic strata, with progressively attenuated effect sizes from T1 to T3 and significant interaction results under both METS-VF and METS-IR stratifications. PAX4, CPEB3, and HHEX also demonstrated stronger associations in metabolically favorable groups, supported by nominally significant interaction and heterogeneity tests. CDKN2A/B showed associations across metabolic strata, with evidence of differences in genetic effects. KCNQ1 exhibited broadly consistent associations, although evidence of interaction varied by metabolic index. CNNM2, HNF1B, HNF4A, SLC30A8, and DNAJC5B–TRIM55 displayed subgroup-dependent patterns of statistical significance, which did not uniformly correspond to significant interaction effects. ALDH2 showed distinct association patterns across metabolic strata and between men and women. These findings suggest that genetic associations with T2DM may vary according to metabolic context, although subgroup-specific genome-wide significance does not necessarily establish differential genetic effects. Formal interaction and heterogeneity analyses provide additional evidence for metabolic effect modification at selected loci, suggesting that metabolic stratification may help characterize variation in genetic associations with T2DM in the Korean population.