A review is given summarizing the present knowledge of bone turnover markers with special emphasis on biological, preanalytical and technical criteria in the proper judgement of efficacy and limitations of the methods employed. The marker substances may be either measures of bone formation or bone resorption. Markers of bone formation are bone alkaline phosphatase, osteocalcin and the carboxyl-terminal propeptide of procollagen type I. Bone alkaline phosphatase has proved to be superior to total alkaline phosphatase activity with respect to diagnostic sensitivity and specificity. Immunochemical techniques for measuring bone alkaline phosphatase show a cross-reactivity of 14-20% with liver alkaline phosphatase. However, this does not compromise the clinical usefulness of these assays except for patients with severe liver diseases. Osteocalcin is strictly bone-specific but shows numerous disadvantages with respect to apparent instability and discordant results as obtained by different methods; however, in certain diagnostic situations (corticosteroid-induced osteopenia, absence of destroyed bone architecture) osteocalcin may serve as a sensitive bone turnover marker. The carboxyl-terminal propeptide of procollagen type I generally shows low discriminating power in the diagnosis of bone diseases. The urinary excretion of pyridinium "cross-links' has been carefully evaluated so far with respect to analytical performance and clinical usefulness. This marker may be a substitute for 4-hydroxyproline measurements as the method of choice for assessment of bone resorption. There are other degradation products from the telopeptide regions of bone-derived collagen type I which are excreted into the urine (N-telopeptides, CrossLapsTM); these analytes are promising tools in the assessment of bone resorption but require further evaluation, in particular with respect to their extraskeletal clearance and putative origin outside bone. Moreover, their clinical usefulness may vary depending on the patient group examined. In contrast, the serum concentration of the cross-linked telopeptide region of collagen type I seems to lack both diagnostic specificity and sensitivity in the majority of patient groups. Tartrate-resistant acid phosphatase (as determined by the presently available methods) cannot be recommended as a routine tool for assessment of bone resorption.