Objective: To assess a wide spectrum of biochemical markers (biomarkers) in a large cohort of individuals with (very) early symptomatic knee and/or hip osteoarthritis (OA). Secondly, to investigate associations between biomarkers and between biomarkers and demographics to demonstrate validity of the obtained dataset and further investigate the involvement and/or role of these biomarkers in OA. Design: Fourteen biomarkers (uCTX-II, uCTX-I, uNTX-I, sCOMP, sPIIANP, sCS846, sC1,2C, sOC, sPINP, sHA, sPIIINP, pLeptin, pAdiponectin, pResistin) were assessed by ELISA or RIA in CHECK (Cohort Hip and Cohort Knee), a 10-year prospective cohort of 1,002 individuals with early symptomatic knee and/or hip OA. Results: Quality controls revealed that gathered data were technically reliable. The majority of biomarkers showed relevant associations with demographic variables, which were expectedly different between genders and/or menopausal status for some. Principal component analysis enabled identification of five clusters, consecutively designated as 'bone-CTX-II', 'inflammation', 'synovium', 'C1,2C-adipokines', and 'cartilage synthesis' cluster. Notably, uCTX-II clustered with biomarkers of bone metabolism, while sCOMP clustered with biomarkers of synovial activity. Conclusions: The identified clusters extended knowledge on individual biomarkers from mostly smaller studies as did the observed associations between biomarker levels and demographics, from which validity of our data was deduced. uCTX-II may not only reflect articular cartilage but also bone metabolism and sCOMP may reflect synovial rather than cartilage metabolism. Major involvement of adipokines in joint metabolism was not identified. © 2012 Osteoarthritis Research Society International.