We examined OPG and soluble RANKL in the serum (sOPG, sRANKL) and synovial fluid (synOPG, synRANKL) in patients with rheumatoid arthritis (RA) and osteoarthritis (OA). OPG and RANKL were measured in 85 patients (44 with RA, 41 patients with OA) in serum and synovial fluid as well. For measuring of OPG and RANKL ELISA tests were used. The results of OPG and RANKL were compared with clinical and radiological scores. We found a negative correlation for OPG and RANKL in synovial fluids: not only for the whole group of patients (P < 0.003, r = -0.32), but also for the subgroups (RA: P < 0.04, r = -0.28, OA: P < 0.002, r = -0.54). SRANKL and synRANKL were positively correlated in the whole group (P < 0.01, r = 0.25) and in the OA group (P < 0.02, r = 0.35); the RA group was showing a trend (P < 0.063, r = 0.24), however. Serum OPG was lower in RA, synOPG higher in OA. The difference between the two patient groups was only significant for synOPG (P < 0.03, r = 0.056), but not for sOPG (P < 0.09, r = 0.19), sRANKL (P < 0.43, r = 0.85) or synRANKL (P < 0.11, r = 0.22). The synOPG:synRANKL ratio was significantly correlated with the Larsen score (P < 0.004, r = 0.38). Synovial OPG is significantly decreased in rheumatoid joints, whereby synovial RANKL is increased. Lower synOPG could reflect a lower protective effect on bone, thus leading to an earlier and more pronounced bone destruction in RA. However, the effect of different mediators for joint destruction in RA and OA seems not to be important to the pathophysiological changes in the joints. The upregulation of serum OPG might be the result of the inflammation; in contrast, an upregulation of RANKL could not be found in the serum of patients with RA and OA.