Background Ultrasonography (US) is increasingly an important tool in diagnosis and management of rheumatoid arthritis (RA). However, the knowledge regarding to what extent US adds information to clinical examination at the individual joint level is limited.

Objectives To assess concordance between clinical joint examination and US in detecting pathology at the individual joint level in a cohort of early RA patients.

Methods 228 patients with early RA who fulfilled the 2010 ACR/EULAR classification criteria were recruited at 11 rheumatology centers between 2010 and 2013. Patients were required to have symptom duration <2 years and to be DMARD naïve with indication for DMARD treatment. 44 joints were assessed for swelling. US examinations were performed by experienced sonographers, using a validated gray-scale (GS) and power-Doppler (PD) semi-quantitative scoring system with ranges 0-3 for GS and PD in each of the following 36 joints: MCP 2-5, radio-carpal (RCJ), distal radio-ulnar (DRUJ), inter-carpal (ICJ), PIP 2-3, elbow, knee, talo-crural and MTP 1-5 bilaterally. We defined US synovitis (US+) as GS ≥2 and/or PD ≥1. For the wrist the comparison with clinical joint swelling included RCJ, DRUJ and ICJ. Joints were assessed bilaterally and analysed on an aggregated level. Kappa values between clinical joint swelling and US synovitis were calculated for all joints. We specifically assessed the proportion of joints with US synovitis that were not clinically swollen, and conversely.

Results The mean age (SD) of the 228 patients was 51.1 (13.8) years, mean (SD) disease duration 7.1 (5.4) months and 62.3% were female. A total of 8208 joints were examined by US. Agreement between US and clinical examination was best in the elbow, wrist and knee (kappa 0.68, 0.53, 0.51, respectively). The proportion of joints with US synovitis that were not clinically swollen ranged from 16% (PIP2) to 76% (MTP1). The three joints with the largest and the three with the smallest discrepancies are presented in the table. Similarly, a significant proportion of clinically swollen joints had no US synovitis, ranging from 23% (wrist and elbow) to 70% (ankle).

Conclusions In this study of patients with early RA we found the level of agreement between US and clinical examination to be somewhat higher in medium to large joints. However, substantial discrepancies between clinical joint swelling and US pathology were found for most joints, and the findings suggest that US provides additional information compared to clinical examination in most joint areas.

Disclosure of Interest L. Nordberg: None declared, E. Lie: None declared, S. Lillegraven: None declared, A.-B. Aga: None declared, I. Olsen: None declared, T. Uhlig: None declared, D. van der Heijde: None declared, T. Kvien: None declared, E. Haavardsholm Grant/research support from: Abbvie, Pfizer, MSD, Roche, UCB.