The relationship between synovial fluid biomarkers and characteristics of pain sensitization in patients with knee osteoarthritis: a preliminary analysis

Aim: Synovitis and central sensitization are both common features in knee osteoarthritis (KOA) patients. However, the relationship between sensitization and synovitis is unclear. Our recent work demonstrated that exercise therapy is effective in the reduction of intra-articular inflammation. Moreover, there is growing evidence that exercise may restore the endogenous analgesia and thereby modulate pain sensitization in KOA patients. Therefore, in the present study, we sought to investigate the relationship between sensitization and intra-articular biomarker levels in KOA patients.

Methods: Seventeen (6M, 11F) KOA patients (age ≥50y) scheduled for total knee arthroplasty in UZ Brussel were included in the study. Pressure pain thresholds (PPTs), temporal summation (TS) and conditioned pain modulation (CPM) were assessed in the patients. During surgery, synovial fluid was obtained to determine biomarker (i.e. CXCL10, CCL2, IL-10, NGF, BDNF, MMP-1, MMP-7, CXCL9) levels by a multiplexed bead-based immunoassay. Multiple linear regression was used to assess the association between pain sensitization parameters (PPT, TS, CPM) and intra-articular biomarker levels. Four independent predictors were created including inflammation (log(CXCL-10*CXCL2/IL-10)), cartilage degeneration (log(MMP-1*MMP-7)), sensitization (log(BDNF*NGF)), and ageing (log(CXCL-9)). Regression models were corrected for BMI, sex and age.

Results: Two out of eight PPT locations around the knee (i.e. 1 and 8) seemed most relevant and were chosen as dependent variables in the regression model, besides the control sites; m. tibialis anterior and m. ext. carpi radialis longus. Log(CXCL-10) was inversely correlated with mean PPT at locations 1 and 8 (Pearson’s r: -0.518 and -0.502, respectively). No regression model could significantly predict the mean PPT at locations 1 and 8 around the knee, the mean PPT at the m. tibialis anterior and m. ext. carpi radialis longus, or the absolute TS. However, 61.8% (R²=0.618) of the variation in relative CPM effect could be significantly explained by the cartilage degeneration predictor, MMP-1*MMP-7 (p=0.022; n=16; β= 0.8).

Conclusion: Top-down pain inhibition can partially be predicted by intra-articular MMP-1 and MMP-7. However, more research is necessary to further elucidate this relationship.