OBJECTIVE: Hemorrhagic arthritis (HA) is a common disease of the musculoskeletal system caused by hemorrhage in the joints, leading to damages in the synovium and cartilage. Pure platelet-rich plasma (P-PRP) has been recently demonstrated to have anti-inflammatory and regenerative potential attributed to the various cytokines and growth factors that it contains. The aim of this study was to investigate the efficacy of P-PRP for the treatment of patients with mild and severe HA. METHODS: Autologous blood was withdrawn from the New Zealand rabbits and injected into their left and right knees to establish the HA models. The injection was performed once a week and repeated for 8 weeks to establish the mild HA models and for 16 weeks to establish the severe HA models. One hundred microliters of P-PRP was injected into the left HA knees, and the same volume of sterile 0.9% saline was injected into the corresponding right knees. MRI examination, H&E staining, and toluidine blue staining were respectively performed to evaluate the histological difference of synovium and cartilage between the P-PRP treated and untreated mild and severe groups. Normal knees were set as the control group. RESULTS: Pathological changes including tissue color, joint effusion, and synovium inflammation in the mild treated group were less severe compared to the other three experimental groups based on gross observation. The difference of joint cavity diameter between the mild treated and untreated groups was 2.67 ± 0.75 mm (95%CI, 1.20-4.14 mm), which was significantly larger than that between the severe treated and untreated groups (1.50 mm ± 0.48, 95%CI, 0.56-2.44 mm) (mean difference in joint cavity, 1.17 ± 0.32 mm; 95%CI, 0.49-1.85 mm; P < 0.01). MRI examination showed the synovitis and bone marrow edema were less severe in the treated groups compared to the corresponding untreated groups. H&E staining of the synovium suggested that the inflammation was much less and the cell number was much smaller in the treated than in the untreated HA knees in regard to both the mild and severe groups. The mean difference of cell number between the mild treated and untreated groups was 307.40 ± 14.23 (95%CI, 241.54-343.26; P < 0.001), which was 699.20 ± 82.80 (95%CI, 508.26-890.14; P < 0.001) between the severe treated and untreated groups. H&E staining and toluidine blue staining of the cartilage exhibited an obvious amelioration of inflammation and cartilage matrix loss after being treated with P-PRP for both groups, especially the mild group. CONCLUSIONS: P-PRP was effective for the treatment of HA by inhibiting the development of synovitis and cartilage matrix loss in the affected joints, which was particularly obvious in the early-stage HA. This study supports the view that there is a great potential of P-PRP to be considered and used as a non-operative treatment for hemorrhagic arthritis at its early stage.