Background Ultrasound, as compared with clincal assessment, is a sensitive tool for evaluating synovitis in patients with rheumatoid arthritis (RA). A semi-quantitative score, gray scale (GSUS) and color doppler (CDUS) mode, has been developed to grade synovitis in RA patients.
Objectives To evaluate the agreement between clinical assessment performed by two independent observers, in patients with RA; to evaluate the inter-observer agreement between ultrasound (US) evaluation of twelve joints, performed by two experienced sonographers; to correlate clinical evaluation with US data and disease duration.
Methods Thirty nine patients with RA were selected from database of a rheumatology unit according to age, sex and treatment, by a blinded observer to the clinical activity and duration of disease. The patients underwent clinical evaluation by an experienced rheumatologist and a final year medical student. The number of swollen (SJ) and tender joints included in DAS 28 score were recorded. Two different experienced sonographers blinded to clinical and demographic data performed US evaluation in gray scale (GSUS), color Doppler (CDUS) and detection of erosions, of twelve joints (bilaterally: wrists, second, third and fifth metacarpophalangeal joints and second and fifth proximal interphalangeal joints). Cohen's Kappa was calculated to determine agreement between clinical and sonographic findings. Spearmen's correlation coefficients were calculated to determine correlations between clinical and sonographic findings.
Results Agreement in clinical evaluation between the two observers was strong: k=0,62 for tender joints and K=0,80 for swollen joints. Agreement of US evaluation between the two sonographers was very strong: K=0,889 in GS, K=0,917 in CD and K=0,854 for detection of erosions. Concerning the study of 468 joints, GSUS and CDUS found respectively 3 and 1,05 more synovitis than SJ count (sonographic criteria for synovitis in GS and CD ≥1). When we considered a sonographic criteria for synovitis GSUS ≥2, we found 2,09 more synovitis than SJ count. There were a significant statistical correlation between disease duration and the total GS evaluation.
Conclusions In this study we found a very strong inter-observer agreement in GSUS, CDUS findings and detection of erosions in a RA population, performed by two experienced sonographers, making this technique as a very reliable tool in daily clinical practice. As we detect much more synovitis in GSUS than SJ count this is favour of the hypothesis that clinical examination in far from optimal for assessing joint synovitis in RA patients. We found different results between GSUS assessment and clinical examination when we considered different cut-offs of synovitis, raising the question how joint activity should be defined.
Gärtner M et al. Sonographic Joint Assessment in Rheumatoid Arthritis. Arthritis Rheum 2013; vol 65 (8): 2005-2014.
Garrigues F et al. Concordance between clinical and ultrasound findings in rheumatoid arthritis. Joint Bone Spine 80(2013): 597-603.
Disclosure of Interest None declared