Although previous studies indicated that hypoglycemic agents could affect bone metabolism, little is known about whether these agents are associated with the risks of osteoporotic fracture in Japanese patents with type 2 diabetes. We examined whether treatments of diabetes, such as insulin administration, sulfonylurea, thiazolidinedione, and metformin, were associated with the presence of vertebral fractures in 494 men and 344 postmenopausal women with type 2 diabetes. We analyzed the relationships between each treatment versus bone turnover markers, bone mineral density (BMD), and the presence of prevalent vertebral fractures. Multiple logistic regression analysis adjusted for age, duration of diabetes, body mass index, serum creatinine, serum C-peptide, and HbA(1c) showed that, in postmenopausal women, treatments with insulin administration or thiazolidinedione were significantly and positively associated with the presence of vertebral fractures [odds ratio (OR) = 2.27, P = 0.012 and OR = 3.38, P = 0.038, respectively], whereas treatment with sulfonylurea was significantly and inversely associated with vertebral fractures (OR = 0.48, P = 0.018). These relationships were still significant after additional adjustment for lumbar BMD. In contrast, no significant relationships between treatments with any agent and the presence of vertebral fractures were found in men. These findings suggest that postmenopausal women treated with insulin or thiazolidinedione have a high risk of vertebral fractures independent of age, body stature, blood glucose level, insulin secretion, or BMD whereas treatment with sulfonylurea is associated with a decreased risk.