We isolated and characterized a peptide fragment corresponding to amino acid sequence 14-28 of human osteocalcin in urine from Paget's disease, and developed a polyclonal antibody reactive to this peptide in urine. We used this antibody to measure urinary fragments of osteocalcin and compared to efficacy of the urinary osteocalcin assay with a serum osteocalcin (sOC) assay (ELISA-Osteo, Cis-Bio International) to monitor the short-term changes in bone turnover in response to alendronate treatment. The synthetic peptide-based urinary osteocalcin (uOC) radioimmunoassay (RIA) showed an analytical sensitivity of 6.25 ng/mL, standard curve range of 3.12-400 ng/mL, and mean intra- (n = 20) and interassay (n = 30) coefficient of variation (CV) of <15%. Urine osteocalcin concentrations in postmenopausal osteoporotic patients were approximately 90% higher than in normal premenopausal controls. Series of 24 h urine and matched serum samples were collected at baseline, 30 days, and 90 days after treatment of postmenopausal osteoporotic patients with daily dose of 10 mg alendronate. We measured urinary osteocalcin (uOc) levels and urinary N-telopeptide (uNTx, Ostex) in urine samples and serum N-telopeptide (sNTx), C-telopeptide (sCTx, Osteometer), serum osteocalcin (sOC) as well as bone-specific alkaline phosphatase (sALP) (Alkphose-B, Metra Biosystems) in serum samples. The percent change data obtained between baseline and 30 days (n = 18) posttreatment suggested a rapid decline in uOC concentration (-27%, p < 0.01) in response to alendronate treatment, as compared with a marginal and nonsignificant decrease in sOC (-7.2%, p = 0.417) or sALP (-3.4%, p = 0.689), two specific markers of bone formation. As expected, due to the coupling of bone formation and bone resorption, the concentration of all markers showed a 30%-45% decline compared with baseline values after 90 days (n = 16) of treatment. Correlation of markers after a 30 day treatment with alendronate revealed a higher correlation (r = 0.61, p < 0.01) between uOC and uNTx, as compared with sOC (r = 0.03, p = 0.447) or sALP (r = -0.14, p = 0.295) with uNTx. Similarly, correlation coefficients with r values between 0.48 and 0.55 (p < 0.05) were observed between uOC, sNTx, and sCTx, whereas no significant correlation was observed between sOC and sNTx or sCTx. These results provide indirect evidence that fragments measured by the urine assay probably originated from bone resorption, and suggest that the uOC assay could be used to assess short-term changes in bone metabolism with regard to osteocalcin.