AMP-activated protein kinase (AMPK) modulates cellular energy levels by regulating glucose transport, fatty acid oxidation and oxidative phosphorylation. AMPK expression rises with exercise/energy depletion, and metformin increases AMPK activity in skeletal muscle and liver. We therefore examined whether common variants in the 7 genes that encode AMPK subunits affect progression from IGT to diabetes (DM) and responses to preventive interventions (lifestyle modification or metformin, vs placebo) in the DPP. We characterized the haplotype structures of the 7 AMPK genes, selected 66 single-marker and 60 multi-marker tests that capture common variation in Caucasian and African HapMap samples, and genotyped them in 2,994 participants. We performed Cox regression analysis using genotype, intervention, and their interactions as predictors of DM incidence, and corrected for the many variants tested after considering linkage disequilibrium (LD) at each locus. Four variants in the AMPK gene PRKAA2 (in moderate LD with each other: r²=0.4-0.9), one variant in PRKAB2 and one variant in PRKAG3 showed nominally significant interactions with the lifestyle intervention (P≤0.01) but not metformin; in all six, the genotypic groups differed by waist circumference and/or BMI at baseline (P≤0.01-0.001), and subjects in the lifestyle arm who carried the genotype associated with the leaner phenotype were less likely to develop DM (P≤0.01). Adjustment for waist circumference did not alter the results. In addition, minor allele homozygotes at four PRKAG1 variants (all in strong LD with each other: r²>0.8) were more likely to develop DM than major allele homozygotes (P≤0.007), with hazard ratios ∼1.75 (95% CI 1.16-2.70) in the placebo arm. There were nominally significant interactions with metformin at 5 PRKAA1, 1 PRKAA2, 2 PRKAB2 and 4 PRKAG2 variants (P=0.007-0.04). None of these associations was statistically significant after correction for multiple hypothesis testing. We conclude that variants in AMPK genes may influence adiposity and affect the response to preventive interventions. Confirmation of these effects in other samples is needed.