Recent genome-wide association studies performed in selected patients and control participants have provided strong support for several new type 2 diabetes susceptibility loci. To get a better estimation of the true risk conferred by these novel loci, we tested a completely unselected population of type 2 diabetes patients from a Norwegian health survey (the HUNT study).

We genotyped single nucleotide polymorphisms (SNPs) in PKN2, IGFBP2, FLJ39370 (also known as C4ORF32), CDKAL1, SLC30A8, CDKN2B, HHEX and FTO using a Norwegian population-based sample of 1,638 patients with type 2 diabetes and 1,858 non-diabetic control participants (the HUNT Study), for all of whom data on BMI, WHR, cholesterol and triacylglycerol levels were available. We used diabetes, measures of obesity and lipid values as phenotypes in case-control and quantitative association study designs.

We replicated the association with type 2 diabetes for rs10811661 in the vicinity of CDKN2B (OR 1.20, 95% CI: 1.06–1.37, p = 0.004), rs9939609 in FTO (OR 1.14, 95% CI: 1.04–1.25, p = 0.006) and rs13266634 in SLC30A8 (OR 1.20, 95% CI: 1.09–1.33, p = 3.9 × 10⁻⁴). We found borderline significant association for the IGFBP2 SNP rs4402960 (OR 1.10, 95% CI: 0.99–1.22). Results for the HHEX SNP (rs1111875) and the CDKAL1 SNP (rs7756992) were non-significant, but the magnitude of effect was similar to previous estimates. We found no support for an association with the less consistently replicated FLJ39370 or PKN2 SNPs. In agreement with previous studies, FTO was most strongly associated with BMI (p = 8.4 × 10⁻⁴).

Our data show that SNPs near IGFBP2, CDKAL1, SLC30A8, CDKN2B, HHEX and FTO are also associated with diabetes in non-selected patients with type 2 diabetes.