We examined whether or not BMD or bone markers were useful for assessing the risk of vertebral fractures in 248 Japanese men with type 2 diabetes. We analyzed the relationships between bone markers (osteocalcin [OC], bone-specific alkaline phosphatase [BAP], urinary N-terminal cross-linked telopeptide of type-I collagen) or BMD and HbA(1c), urinary C-peptide, insulin-like growth factor-I (IGF-I), parathyroid hormone, 1,25(OH)(2) vitamin D, and the presence of prevalent vertebral fractures. Multiple regression analysis adjusted for age, body height, weight, duration of diabetes, and serum creatinine showed that serum OC and OC/BAP ratio were correlated negatively with HbA(1c) (P < 0.01) and positively with IGF-I (P < 0.01). Multivariate logistic regression analysis adjusted for the above parameters showed that serum OC/BAP ratio was inversely associated with the presence of vertebral fractures (odds ratio = 0.695, P < 0.05). This association was still significant after additional adjustment for lumbar or femoral neck BMD. Our results suggest that poor diabetic control and lower IGF-I level are linked to impaired bone formation and resultant reduction in OC/BAP ratio in men with type 2 diabetes. The OC/BAP ratio could be clinically useful for assessing the risk of vertebral fractures independent of BMD in diabetic men.