Comparison of Bone Turnover Biomarkers in Serum and Urine Measured on an Automated Analytical Platform
Abstract Background Matched serum and urine samples from patients who had total hip replacement were used to assess serum-validated immunoassay reagents for use in urine. Methods Samples were evaluated by an automated electrochemiluminescent immunoassay (cobas e411; Roche Diagnostics) for C-terminal telopeptide of type I collagen isoform β (β-Crosslaps), osteocalcin N-terminal midfragment (N-MID OC), N-terminal propeptide of type I collagen (PINP), and interleukin 6 (IL-6). Spike and recovery experiments were utilized to assess urinary matrix effects. Correlations between serum and both raw and creatinine-corrected urinary measures were assessed. Short-term precision was assessed. Results Spike and recovery experiments indicated minimal matrix effects of urine for the β-Crosslaps assay. Potential matrix effects were observed for the other analytes because N-MID OC and IL-6 tended to be slightly overrecovered, whereas PINP was underrecovered. There were strong correlations between serum β-Crosslaps and raw (Spearman ρ [rs] = 0.725, P < 0.0001) and creatinine-corrected (rs = 0.793, P < 0.0001) urinary measures and moderate correlations between serum N-MID OC and raw (rs = 0.582, P < 0.0001) and creatinine-corrected (rs = 0.482, P < 0.0001) urinary measures. PINP was not detected in urine, and no significant serum–urine correlations were found for IL-6. Short-term precision for urinary levels of β-Crosslaps, N-MID OC, and IL-6 were 1.6%, 6.3% and 14.4%, respectively. Conclusions Urinary measurements of β-Crosslaps and N-MID OC assays were correlated with serum measurements and had good short-term precision. Urinary PINP was not detectable. IL-6 can be measured in urine using this technology, but the levels did not correlate with serum levels, and the short-term precision was variable.