Diurnal variation of serum and urine biomarkers in patients with radiographic knee osteoarthritis

Objective. To evaluate diurnal variation of biomarkers in subjects with osteoarthritis (OA) of the knee. Methods. Twenty subjects with radiographic knee OA were admitted to the General Clinical Research Center of Duke University for an overnight stay to undergo serial blood and urine sampling. Biomarkers measured included serum hyaluronan (HA), cartilage oligomeric matrix protein (COMP), keratan sulfate (KS-5D4), aggrecan neoepitope (CS846), high-sensitivity C-reactive protein (hsCRP), osteocalcin, transforming growth factor beta 1 (TGF beta 1), and type II collagen (CII)-related epitopes (neoepitope from cleavage of CII [C2C], carboxy-terminus of three-quarter peptide from cleavage of CI and CII [C1,2C], and type II procollagen carboxy-propeptide [CPII] in serum, and C-terminal telopeptides of CII [CTX-II] and C2C in urine). Results. Levels of serum RA, COMP, KS-5D4, and TGF beta 1 increased significantly from TO (before arising from bed) to T1 (1 hour after arising). More diurnal variation in HA was observed in patients with higher daily mean HA concentrations. CPII increased significantly from T0 to T2 (4 hours after arising). Urinary concentrations of CTX-II were also found to vary with morning activity, decreasing significantly from T0 to T2. Urinary C2C concentrations increased significantly from T0 until T3 (early evening). No diurnal variations in CS846, hsCRP, osteocalcin, serum C2C, or C1,2C were observed. Six biomarkers (serum C2C, C1,2C, COMP, KS-5D4, TGF beta 1, and urinary CTX-II),were associated with radiographic knee OA (expressed as the sum of Kellgren/Lawrence radiographic severity grades), with the strongest correlations observed with measurements obtained at later time points (either T2 or T3). Conclusion. Our study results suggest that serum and urine sampling for HA, COMP, KS-5D4, TGF beta 1, CPII, urinary CTX-II, and urinary C2C should be standardized in future OA clinical trials. Serum and urine sampling at late midday time points may be the optimal approach for OA studies, although this result should be validated in a larger cohort.