Abstract
Objective: Platelet-rich plasma (PRP) therapy has become a popular treatment for knee osteoarthritis. We aimed to determine the outcomes of knee osteoarthritis patients following PRP therapy using magnetic resonance imaging (MRI) findings and patient-reported outcome measures (PROMs). Design: In this retrospective observational cohort study, we enrolled 161 patients (221 knees) with varus knee osteoarthritis who received multiple PRP injections at our hospital from June 2017 to June 2019. Patients underwent whole-body MRI before and 6 months after treatment. Whole-organ MRI score (WORMS) cartilage integrity and synovial fluid volume were assessed for the medial femorotibial (MFTJ), lateral femorotibial (LFTJ), and patellofemoral joints (PFJ). Pain visual analog scale and Knee Injury and Osteoarthritis Outcome scores were used as PROMs. In addition, a historical control group of 30 patients with medial knee osteoarthritis who did not receive intra-articular injections was evaluated by MRI over the same period for comparison. Results: After 6 months of PRP therapy, the mean WORMS cartilage score of the LFTJ and PFJ and the total WORMS cartilage score for all three joints improved significantly, and synovial fluid volume reduced significantly. Moreover, a reduction in synovial fluid volume correlated with improvements in several KOOS subscales but not with VAS, which may explain the lack of association with responder status. These results suggest that synovial fluid reduction reflects functional improvement but is not a direct surrogate for pain relief. In addition, the change score of WORMS PFJ cartilage correlated positively with clinical outcomes in responders. By contrast, in the control group, no compartment demonstrated improvement in WORMS cartilage scores, and several compartments showed a trend toward deterioration. Conclusions: In this retrospective observational study, PRP therapy was associated with improvements in WORMS cartilage integrity scores and reductions in synovial fluid volume, with partial correlations to patient-reported outcomes. The inclusion of a historical control group strengthens the interpretation of these findings, although definitive conclusions cannot be drawn. Further randomized controlled trials are needed to confirm these preliminary observations.