Objective:Much of the genetic variation in glucose levels remains to be discovered. Especially, research on gene-environment interactions is scarce. Overweight is one of the main risk factors for hyperglycemia. As transcriptional regulation is important for both weight maintenance and glucose control, we analyzed 353 single nucleotide polymorphisms (SNPs), occurring in transcriptional pathways of glucose and lipid metabolism in interaction with body mass index (BMI) on glucose levels.Research design and methods:SNPs were measured in 3244 participants of the Doetichem cohort. Non-fasting glucose levels and BMI were measured twice in 6 years. SNP x BMI interactions were analyzed by mixed models and adjusted for age, sex, time since last meal, and follow-up time. False discovery rate (FDR) <0.2 was used to adjust for multiple testing.Results:Two SNPs in the PPARGC1A gene (rs8192678, FDR=0.07; rs3755863, FDR=0.17) showed a significant interaction with BMI. The rare allele of both SNPs was associated with significantly lower glucose levels in subjects with a BMI</=25 kg m(-2) (rs8192678, P=0.02; rs3755863, P=0.03). An inverse association was suggested in subjects with a BMI>28 kg m(-2). A small intervention study (n=120) showed similar, though non-significant, results.Conclusions:Using a pathway-based approach, we found that BMI significantly modified the association between two SNPs in the PPARGC1A gene and glucose levels. The association between glucose and PPARGC1A was only present in lean subjects. This suggests that the effect of the PPARGC1A gene, which is involved both in fatty acid oxidation and glucose metabolism, is modified by BMI.International Journal of Obesity advance online publication, 2 February 2010; doi:10.1038/ijo.2009.302.