OBJECTIVE: To determine the relationship between biochemical bone markers and knee cartilage volume and cartilage loss over 2 years; and to investigate whether bone markers are useful to predict the cartilage loss in healthy men. METHODS: Forty healthy Caucasian men (mean age 52.3 yrs) with no symptoms of osteoarthritis (OA) were recruited. Each subject had magnetic resonance imaging (MRI) performed on his dominant knee at baseline and 2 years later. Serum level of osteocalcin (OC), urinary levels of pyridinoline (PYD) and deoxypyridinoline (DPD), and total body bone mineral content (BMC) were measured at baseline. Tibial plateau bone size was measured at baseline. Tibial cartilage volume was measured at baseline and at followup, by means of image processing. RESULTS: Twenty-eight men (70%) completed the longitudinal MRI component of the study. At baseline, no significant associations were observed between values of serum OC or urine PYD and DPD and tibial cartilage volume. Higher baseline serum OC level tended to be associated with a decreased rate of cartilage loss (p = 0.06); no significant association was shown between baseline urine PYD and DPD and tibial cartilage loss, after adjusting for age, body mass index, total body BMC, and tibial plateau bone size. CONCLUSION: Higher baseline serum OC level tended to be associated with a decreased rate of cartilage loss, suggesting that increased bone formation may protect against tibial cartilage loss over 2 years. Studies are needed to determine the role of bone metabolism in the pathogenesis of knee OA.