A Multicenter Study to Evaluate Harmonization of Assays for C-Terminal Telopeptides of Type I Collagen (ß-CTX):A Report from the IFCC-IOF Committee for Bone Metabolism (C-BM)

Background: Biochemical bone turnover markers are useful tools to assess bone remodeling. C-terminal telopeptide of type I collagen (ß-CTX) has been recommended as a reference marker for bone resorption in research studies. Methods: We describe the results of a multicenter study for routine clinical laboratory assays for ß-CTX in serum and plasma. Four centers (Athens GR, Copenhagen DK, Liege BE and Sheffield UK) collected serum and plasma (EDTA) samples from 796 patients presenting to osteoporosis clinics. Specimens were analyzed in duplicate with each of the available routine clinical laboratory methods according to the manufacturers’ instructions. Passing-Bablok regressions, Bland–Altman plots, V-shape evaluation method, and Concordance correlation coefficient for ß-CTX values between serum and plasma specimens and between methods were used to determine the agreement between results. A generalized linear model was employed to identify possible variables that affected the relationship between the methods. Two pools of serum were finally prepared and sent to the four centers to be measured in 5-plicates on 5 consecutive days with the different methods. Results: We identified significant variations between methods and between centers although comparison results were generally more consistent in plasma compared to serum. We developed univariate linear regression equations to predict Roche Elecsys®, IDS-iSYS, or IDS ELISA ß-CTX results from any other assay and a multivariable model including the site of analysis, the age, and weight of the patient. The coefficients of determination (R2) increased from approximately 0.80 in the univariate model to approximately 0.90 in the multivariable one, with the site of analysis being the major contributing factor. Results observed on the pools also suggest that long-term storage could explain the difference observed with the different methods on serum. Conclusion: Our results show large within- and between-assay var