Fracture-induced changes in bone turnover: a potential confounder in the use of biochemical markers in osteoporosis