PURPOSE OF THE STUDYA failure of total hip or knee artroplasty is associated with an increased production of joint fluid. This contains wear particles and host cells and proteins, and is assumed to be involved in the pathogenesis of aseptic loosening and periprosthetic osteolysis. This study investigated the effect of synovial fluid from patients with aseptically failed joint prostheses on osteoblast cultures. MATERIAL AND METHODSSynovial fluid samples were obtained from patients with failed total joint prostheses (TJP; n=36) and from control patient groups (n=16) involving cases without TJP and osteoarthritis, without TJP but with osteoarthritis, and with stable TJP. The samples were treated in the standard manner and then cultured with the SaOS-2 cell line which shows the characteristics and behaviour of osteoblasts. Each fluid sample was also examined for the content of proteins, cells and selected cytokines (IL-1β, TNF-α, IL-6, RANKL and OPG detected by ELISA). We tested the hypothesis assuming that the fluids from failed joints would show higher cytotoxicity to osteoblast culture and we also expected higher levels of IL-1β, TNF-α, IL-6, and RANKL in patients with TJP failure and/ or with more severe bone loss. The statistical methods used included the Kruskal-Wallis ANOVA and Mann-Whitney U test. RESULTSThe fluids from failed TJPs showed the highest RANKL and the lowest OPG levels resulting in the highest RANKL/OPG ratio. However, there was no evidence suggesting that the joint fluids from failed TJPs would be more toxic to osteoblast culture than the fluids from control groups. In addition, no correlation was found between the fluid levels of molecules promoting inflammation and osteoclastic activity and the extent of bone loss in the hip (in terms of Saleh's classification) or the knee (AORI classification). In fact, the fluids from failed TJPs had higher protein levels in comparison with the controls, but the difference was not significant. DISCUSSIONThe finding of high RANKL levels and low OPG concentrations is in agreement with the theory of aseptic loosening and periprosthetic osteolysis. The other cytokines, particularly TNF-α and IL-1β, were found in low levels. This can be explained by the stage of particle disease at which the samples were taken for ELISA analysis. It is probable that the level of signal molecules reflects osteolytic process activity and is therefore not constant. The reason for no correlation found between cytokine levels and the extent of bone loss may also lie in the use of therapeutic classifications of bone defects that is apparently less sensitive to the biological activity of aseptic loosening and/or periprosthetic osteolysis. CONCLUSIONSSynovial fluids from failed total hip or knee joint prostheses are not toxic to osteoblast cultures. Cytotoxicity indicators and levels of pro-inflammatory and pro-osteoclastic cytokines (IL-1β, TNF-α, IL-6, RANKL and OPG) do not correlate well with the extent of periprosthetic bone loss.