Whether biomarkers associated with cardiovascular disease risk also predict incident diabetes mellitus (DM) is unknown. Our objective was to determine if a panel of 18 biomarkers previously associated with risk of cardiovascular disease also predicts incident DM in statin-treated patients with coronary artery disease (CAD). The Treating to New Targets (TNT) study is a randomized trial that compared the efficacy of high (80 mg) versus low (10 mg) dose atorvastatin for the secondary prevention of coronary heart disease events. Fasting plasma levels of standard lipids and of 18 emerging CAD risk biomarkers were obtained after an 8-week run-in period on atorvastatin 10 mg in a random sample of 1,424 TNT patients. After exclusion of patients with DM at baseline (n = 253), 101 patients developed DM during the median follow-up of 4.9 years. Patients with incident DM had lower levels of total and high-molecular weight adiponectin, lipoprotein-associated phospholipase A2 (Lp-PLA2), soluble receptor of advanced glycation end products, and vitamin D compared with patients without incident DM. In contrast, insulin, soluble CD40 ligand, and soluble intercellular adhesion molecule-1 levels were higher in patients with incident DM compared with those without. Plasma levels of C-reactive protein, cystatin C, lipoprotein(a), monocyte chemotactic protein-1, matrix metalloproteinase-9, myeloperoxidase, neopterin, N-terminal fragment of pro–B-type natriuretic peptide, osteopontin, and soluble vascular cell adhesion molecule-1 were comparable in patients with and without incident DM. After multivariate adjustment, total and high-molecular weight adiponectin as well as Lp-PLA2 were negatively associated with incident DM. Results of this study suggest that plasma lipids and some emerging CAD risk biomarkers, such as adiponectin and Lp-PLA2, may be useful for predicting incident DM in statin-treated patients with stable CAD.