Synovial Fluid Cell Profiling Identifies Distinct Inflammatory Phenotypes in Late‐Onset Rheumatoid Arthritis and Polymyalgia Rheumatica

Objective Late‐onset rheumatoid arthritis (LORA) and polymyalgia rheumatica (PMR) can be difficult to distinguish, particularly in anti–citrullinated protein antibody (ACPA)‐negative patients. We aimed to characterize synovial fluid immune cell profiles in ACPA‐positive LORA, ACPA‐negative LORA, and PMR, and to explore biologically informed patient stratification using an unbiased approach. Methods Synovial fluid was aspirated from inflamed periarticular synovial structures of the shoulder, including the subacromial and subdeltoid bursae and the tendon sheath of the long head of the biceps, in 21 patients with ACPA‐positive LORA (n = 5), ACPA‐negative LORA (n = 8), or PMR (n = 8). Immune cell composition was analyzed by mass cytometry. Principal component analysis and hierarchical clustering were performed to stratify patients. Results ACPA‐positive LORA demonstrated neutrophil predominance compared with PMR. In contrast, PMR demonstrated a neutrophil‐poor profile characterized by higher proportions of T cells and natural killer cells. ACPA‐negative LORA showed substantial heterogeneity, with individual patients displaying either neutrophil‐rich or neutrophil‐poor patterns. Unsupervised clustering of mass cytometry data segregated ACPA‐positive LORA and PMR into distinct clusters, whereas ACPA‐negative LORA cases were distributed across both clusters. Conclusion This exploratory study suggests that late‐onset inflammatory arthritis comprises at least two distinct synovial inflammatory phenotypes that may not fully correspond to current clinical diagnoses. ACPA‐negative LORA appears to represent a heterogeneous condition encompassing these distinct states of inflammation. Synovial fluid immunophenotyping may complement conventional classification and support biologically informed therapeutic stratification.