Bone remodeling may be involved in the pathogenesis of stress fractures in athletes. We conducted a 12-month prospective study to evaluate bone turnover in 46 female and 49 male track and field athletes aged 17-26 years (mean age 20.3; SD 2.0) 20 of whom developed a stress fracture. Baseline levels of bone turnover were evaluated in all athletes and monthly bone turnover levels were evaluated in a subset consisting of the 20 athletes who sustained a stress fracture and a matched comparison group who did not sustain a stress fracture. Bone formation was assessed using serum osteocalcin (OC) measured by human immunoradiometric assay and bone resorption by urinary excretion of pyridinium cross-links (Pyr and D-Pyr); high performance liquid chromatography and N-telopeptides of type 1 collagen (NTx) using ELISA assay. Athletes who developed stress fractures had similar baseline levels of bone turnover compared with their nonstress fracture counterparts (P > 0.10). Results of serial measurements showed no differences in average levels of Pyr, D-Pyr, or OC in those who developed stress fractures (P = 0.10) compared with the control group. In the athletes with stress fractures, there was also no difference in bone turnover levels prior to or following the onset of bony pain. Our results show that single and multiple measurements of bone turnover are not clinically useful in predicting the likelihood of stress fractures in athletes. Furthermore, there were no consistent temporal changes in bone turnover associated with stress fracture development. However, our results do not negate the possible pathogenetic role of local changes in bone remodeling at stress fracture sites, given the high biological variability of bone turnover markers and the fact that levels of bone turnover reflect the integration of all bone remodeling throughout the skeleton.