Background/Purpose: Rheumatoid arthritis (RA) is characterized by considerable synovial inflammation which may result in secondary osteoarthritis (sOA). Primary OA (pOA) occurs in patients without predisposing disorders such as RA and displays varying degrees of synovial inflammation. Although it is conceivable that a major structural difference between sOA due to RA and pOA consists in synovial histology, direct research on such histological differences is limited. Moreover, the search for suitable biomarkers to distinguish OA subtypes is considered an important step in the search for medical treatments. We therefore investigated if synovitis in patients with sOA due to RA is different from pOA, and if it more closely resembles active RA.

Methods: Synovial tissue was collected and snap frozen at time of joint replacement in patients with (1) pOA (n = 8, hip), (2) sOA (defined as RA according to ACR/EULAR criteria plus OA in accordance with 2009 EULAR recommendations for diagnosis of knee OA (n = 4) or 1991 ACR criteria for hip OA (n = 3)), and (3) active RA without any signs of OA (n = 8 metacarpophalangeal joints; mean DAS28 5.2 ± 1.4) by arthroscopically guided biopsy or open synovectomy. Hematoxylin and eosin stained sections were used for determination of the Synovitis Score according to Krenn. Immunohistochemically stained sections were used for semiquantitative scoring or digital image analysis (% stained area) of CD68+ macrophages in a blinded fashion. Comparison of groups was by Kruskal-Wallis Test and Dunn’s posthoc test. Correlations between methods of CD68 scoring were according to Spearman.

Results: High-grade synovitis was revealed by conventional histology in patients with sOA and RA (median scores: 6), but not in pOA where mild synovitis was predominant (median score 3; p = 0.012). This was largely due to the subscore on the inflammatory infiltrate with median scores of 3, 2, and 1 in sOA, RA, and pOA, respectively (p = 0.006), rather than subscores for lining layer hypertrophy (2, 2, 2.5; p = 0.013) or density of resident cells (1, 1.5, 2; p = 0.52). There was good agreement between semiquantitative scoring and digital image analysis for CD68 in all compartments (Spearman’s R 0.85, 0.8, and 0.7 with p < 0.001 for lining, sublining, and total CD68 scoring, respectively). Significant differences between groups were observed in all compartments with the strongest effect in sublining CD68 staining (median stained area 35.6% vs. 15.3% and 2.9% in sOA, RA, and pOA; p = 0.0031). The difference to sOA in sublining CD68 staining was significant after posthoc analysis for pOA, but not RA. Receiver Operating Characteristic analysis confirmed sublining CD68 staining as an excellent marker to distinguish sOA due to RA from pOA (p = 0.026, AUC 0.96, LR 8).

Conclusion: Synovial tissue analysis reveals considerable inflammation in sOA due to RA. Both conventional histology and CD68 staining confirm that sOA is histologically distinct from pOA and more closely resembles active RA. Sublining CD68 is a suitable biomarker to distinguish both OA entities.

ACR Meeting Abstracts - https://acrabstracts.org/abstract/synovitis-in-secondary-osteoarthritis-due-to-rheumatoid-arthritis-a-proof-of-concept-study/