Background Acute knee joint injury is the single greatest risk factor for development of knee osteoarthritis, but only half of these individuals will develop symptomatic disease. Identifying biomarkers of an individual's joint injury response at the outset which are associated with patient relevant outcomes over both the short and long term may allow us to stratify and target therapies at those at highest risk. In a surgical mouse model, where meniscal destabilisation leads to osteoarthritis, we previously identified several molecules which are strongly upregulated in the joint tissues immediately after knee injury, 7 of which were likely to be secreted and detectable in human synovial fluids in clinical translational studies.
Objectives We investigated whether molecules which were up-regulated within hours of surgical joint destabilisation in the mouse were also elevated in the analogous human setting of acute knee injury, how this molecular response varied between individuals, and whether it related to patient-reported outcome measures in the 3 months following injury.
Methods 150 participants were recruited to the Knee Injury Cohort @ Kennedy (KICK), who had recent structural knee injury (<8 weeks from injury). Blood and synovial fluid (SF) at baseline visit, and where available, at 14 days and 3 months following baseline visit were analysed for the 7 candidate molecules. Knee injury and Osteoarthritis Outcome Score (KOOS) was collected at baseline and 3 months, and the summary KOOS4 score analysed. Assays were by MesoScale Discovery™platform or ELISA, and compared for blood and synovial fluid with uninjured control samples.
Results Median age of KICK participants was 25, median time from injury to baseline visit was 17 days and there was substantial pain and impairment by KOOS at baseline visit. 6/7 molecules were significantly elevated in human synovial fluid immediately after injury: IL-6, MCP-1, MMP-3, TIMP-1, activin A and TSG-6. 3/6 molecules were significantly associated with baseline KOOS4 (those individuals with higher SF IL-6, TIMP-1 or TSG-6 had lower KOOS4). These 3 molecules, MMP-3 and activin A were all significantly associated with greater improvement in KOOS4 over 3 months, adjusting for relevant, pre-defined factors including time from injury to sampling, extent of injury, age and presence of heavy blood staining of synovial fluid. When these 5 synovial fluid biomarkers were included in a linear regression model, only synovial fluid IL-6 was independently associated with either baseline KOOS4 (Coeff.-4.0 (-5.92,-2.08), P<0.001), or difference in KOOS4 over 3 months (Coeff.2.80 (0.89,5.52), P=0.043).
Conclusions Our findings validate relevant human biomarkers of joint injury identified from a mouse model. The response, represented best by synovial fluid IL-6, is associated with clinically-relevant outcomes over this early period: a greater inflammatory response soon after the injury is associated with increased impairment and pain by KOOS4 at that time, but does not appear to be an adverse prognostic factor, at least in the short-term, and in this predominantly surgically-managed cohort. Longitudinal outcomes will determine if these molecules are biomarkers of subsequent disease risk.
Disclosure of Interest None declared