Tender joints in early autoantibody-negative rheumatoid arthritis: Should they be included in the scoring of joint involvement of the 2010 ACR/EULAR classification criteria?

Objectives: The interpretation of joint tenderness as a sign of inflammation in patients with autoantibody-negative rheumatoid arthritis (RA) is uncertain. This may hinder disease classification and create selection bias for patient enrollment in clinical trials. Here we tested whether reclassifying the pattern of joint involvement based on swollen joints increases specificity for persistent arthritis in autoantibody-negative early RA. Methods: From a prospective early arthritis cohort in the years 2005-2018, all autoantibody-negative patients fulfilling the 2010 ACR/EULAR RA criteria at enrollment were included. Patients were re-classified for the score of swollen joint involvement (1-3=score 2; 4-10=score 3; >10=score 5). Groups were compared for baseline clinical and ultrasonographic (US) characteristics and outcomes after 12 and 36 months. Results: Of a total of 354 autoantibody-negative patients with 2010-based RA, 39.5 % had a score of swollen joints=5, 47.5 % score=3, and 13 % score=2. We found equal signs of US synovitis and power Doppler of the wrists and metacarpophalangeal joints. Patients with lower swollen joint scores had similar requirements of treatment intensification within month 12 compared with patients with higher baseline inflammation. These latter had the most favourable outcomes, with lower need of second-line treatment strategies within month 36. Exclusion of patients with concomitant fibromyalgia did not modify the results. Conclusions: Joint tenderness should be included in the evaluation of the pattern of joint involvement of the 2010 ACR/EULAR criteria to correctly classify patients with autoantibody-negative early RA. A score solely based on joint swelling may lead to the erroneous under-selection of patients with persistent arthritis.