Clinical application of synovial biopsy in noninflammatory and persistent inflammatory refractory rheumatoid arthritis

Objectives: We aimed to characterise the synovial pathology of refractory rheumatoid arthritis (RA) by comparing two hypothesised clinical phenotypes—persistent inflammatory refractory RA (PIRRA) and noninflammatory refractory RA (NIRRA)—against nonrefractory RA (NORRA). Methods: We conducted a prospective, observational cohort study at two academic rheumatology centres. Adult patients with established RA and active disease (Clinical Disease Activity Index >10) underwent ultrasound-guided synovial tissue biopsies in accordance with European Alliance of Associations for Rheumatology–Outcome Measures in Rheumatology guidelines. Based on the Physician Global Assessment (>2/10) and C-reactive protein (>5 mg/L) thresholds, biological and targeted synthetic disease-modifying antirheumatic drugs (b/tsDMARDs)-inadequate responders patients were classified as PIRRA or NIRRA, whereas b/tsDMARD-naïve were designated NORRA. Histopathological assessments included the Krenn Synovitis Score (KSS) and immunohistochemistry pathotype evaluation (lympho-myeloid, diffuse-myeloid, pauci-immune/fibroid). Results: Of the 93 biopsied patients, 43 were PIRRA, 21 NIRRA, and 29 NORRA. NIRRA had lower KSS (P = .012), lymphoid aggregates (P < .001) and predominantly pauci-immune/fibroid pathotype (n = 21/43, 47.6%), whereas PIRRA displayed more lympho- and diffuse-myeloid pathotypes, coupled with higher inflammatory markers and ultrasound-power Doppler scores. The relative risk of having pauci-immune fibroid synovitis for NIRRA was 1.6 (95% CI 1.2 to 2.9, P = .006), indicating a statistically significant increase in risk compared to PIRRA. Despite having similar overall disease activity scores, NIRRA patients reported significantly greater pain (P = .005) and higher opioid use (P = .015) than PIRRA, and worse health-related quality of life than PIRRA or NORRA, underscoring noninflammatory mechanisms. Conclusions: Our findings demonstrate distinct synovial and clinical phenotypes in refractory RA. Although PIRRA appears driven by active synovial inflammation and immune cell infiltration, NIRRA involves predominantly pauci-immune/fibroid histology with significant noninflammatory contributors to disease burden.