In the case-control analysis, PPARG rs1801282 (Pro12Ala) (OR=1.48 (1.02-2.16)), ADIPOQ rs1063539 (OR=1.17 (1.01-1.35)), and HNF4A rs1884614 (OR=1.16 (1.00-1.32) were associated with T2D (P(allelic)<0.05).
Although no interactions between the four variants on the risk of type 2 diabetes were detected, the multiplicative interaction between PPARG Pro12Ala and HNF4A rs2144908 was found to be associated with 2-hour postprandial insulin (P = 0.004 under an additive model for rs2144908; and P = 0.001 under a dominant model for rs2144908) after adjusting for age, sex and BMI, assuming a dominant model for PPARG Pro12Ala.
Amongst non-obese individuals, we observed significant T2D associations with HNF1A I27L [odds ratio (OR) = 1.14, P = 0.04], GCK -30G>A (OR = 1.23, P = 0.01), SLC30A8 R325W (OR = 0.87, P = 0.04), and TCF7L2 rs7903146 (OR = 1.89, P = 4.5 x 10-23), and non-significant associations with PPARG Pro12Ala (OR = 0.85, P = 0.14), ADIPOQ -11,377C>G (OR = 1.00, P = 0.97) and ENPP1 K121Q (OR = 0.99, P = 0.94).
These variants may contribute significantly to the risk type 2 diabetes conferring insulin resistance of liver, muscle and fat (Pro12Ala) and a relative insulin secretory deficiency (Glu23Lys).