Abstract
Establishing disease biomarkers has been a long-sought after goal to improve Osteoarthritis (OA) diagnosis, prognosis, clinical and pharmaceutical interventions. Given the role of the synovium in contributing to OA, a meta-analysis was performed to determine significant synovial biomarkers in human OA tissue, compared to non-OA patients. Outcomes will direct future research on marker panels for OA disease modelling in vitro/in vivo, aiding clinical research into OA disease targets.
A PRISMA compliant search of databases was performed to identify potential biomarker studies analysing human, OA, synovial samples compared to non-OA/healthy participants. The Risk of Bias In Non-Randomised Studies of Interventions (ROBINS-I) tool assessed methodological quality, with outcome analysed by Grading of Recommendations Assessment, Development and Evaluation (GRADE). Meta-analyses were conducted for individual biomarkers using fixed or random effect models, as appropriate. Where three or more studies included a specific biomarker, Forest Plot comparisons were generated.
3230 studies were screened, resulting in 34 studies encompassing 25 potential biomarkers (1581 OA patients and 695 controls). Significant outcomes were identified for thirteen comparisons. Eleven favoured OA (IL-6, IL-10, IL-13, IP-10, IL-8, CCL4, CCL5, PIICP, TIMP1, Leptin and VEGF), two favoured non-OA controls (BMP-2 and HA). Notably, PIICP showed the largest effect (SMD 6.11 [3.50, 8.72], p <0.00001, I2 99%), and TIMP1 resulted critically important (0.95 [0.65, 1.25], p <0.00001, I2 82%). Leptin and CCL4 showed lower effects (SMD 0.81 [0.33, 1.28], p =0.0009; 0.59 [0.32, 0.86], p <0.0001, respectively).
Thirteen significant synovial biomarkers showed links with OA bioprocesses including collagen turnover, inflammatory mediators and ECM components. Limitations arose due to bias risk from incomplete or missing data, publication bias of inconclusive results, and confounding factors from patient criteria. These findings suggest markers of potential clinical viability for OA diagnosis and prognosis that could be correlated with specific disease stages.
