Bone metastases strongly affect skeletal metabolism by both their growth and their paracrine activities. However, so far no specific laboratory marker has been found to signal the spread of neoplastic disease to bone. We performed a cross-sectional study of 153 cancer patients and an equal number of healthy controls matched for sex and age, in which we determined serum levels of calcium and total alkaline phosphatase (TAP) as well as the fasting urinary excretion of calcium (uCa) and of the collagen cross-links pyridinoline (uPYD) and deoxypyridinoline (uDPD). The aim of the study was to analyze the diagnostic validity of the biochemical parameters measured with regard to neoplastic bone involvement. In the cancer group, 98 patients had overt bone metastases, as judged from radiographic and radioisotopic bone imaging. The remaining 55 patients were also in an advanced stage of disease, but there was no evidence of malignant bone involvement. In comparison to healthy controls, patients both with and without metastatic bone disease had significantly higher levels of TAP, uPYD, and uDPD (P < 0.0001). Only in cancer patients with bone metastases was the median serum calcium level higher than in the healthy controls (P < 0.02). uCa was the same in cancer patients and the control group. Within the collective of cancer patients, individuals with skeletal metastases had higher levels of serum calcium (P < 0.05), TAP (P < 0.01), and uPYD and uDPD (both P < 0.0001), than patients without evidence of malignant bone disease. uCa did not differ between the 2 groups of cancer patients. The cancer patients were then stratified into 4 subgroups according to the serum calcium level (less than or equal to 2.6 mmol/L>) and the absence or evidence of bone metastases. This stratification revealed that in patients with bone metastases, uPYD and uDPD levels were similar in normocalcemic and hypercalcemic subjects, whereas in hypercalcemic patients, uCa levels significantly exceeded those in normocalcemic patients. When the efficacy of TAP, uCa, uPYD, and uDPD in discriminating between patients with and without bone metastases was evaluated by use of receiver-operating characteristic curves and stepwise multivariate regression analysis, uPYD was found to have the highest diagnostic validity. Using 50 mu mol PYD/mol creatinine (i.e. the upper limit of normal range) as the cut-off level, the sensitivity of uPYD was 88.7%, whereas the specificity was only 41.8% (odds ratio, 5.598; 95% confidence interval, 2.547-12.306). Our results suggest that markers of bone formation (TAP) and bone resorption (urinary pyridinium cross-links) differ significantly between patients with and without bone metastases. However, the low specificity of the parameters examined substantially reduces their diagnostic value with regard to neoplastic bone disease. In a further step, longitudinal studies should be performed to evaluate whether these parameters might be useful in monitoring the progression or remission of neoplastic bone disease.
